Using Taxation to Address Noncommunicable Diseases: Lessons From Tonga July 2019 Regional Vice President: Victoria Kwakwa Country Director: Michel Kerf Global Practice Director: Muhammad Ali Pate Practice Manager: Enis Baris Primary Author: Sutayut Osornprasop Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table of Contents EXECUTIVE SUMMARY 14 INTRODUCTION 22 THE COUNTRY CONTEXT 24 THE NCD CRISIS AND DEVELOPMENT IMPACTS 25 NCD RISK FACTORS 26 Tobacco consumption 28 Unhealthy diet 30 Physical Inactivity 33 Harmful alcohol use 34 Obesity 35 THE GOVERNMENT’S NCD RESPONSE 36 MAKING THE MOST OF ‘SIN TAXES’ 40 METHODOLOGY 41 HOUSEHOLD SURVEYS 42 Sampling Design 42 Household characteristics of baseline data 44 Household characteristics of endline data 44 Data analysis 46 RETAIL SURVEYS 46 QUALITATIVE STUDY 46 Focus group discussions 47 Training of fieldwork staff and Pre-test of the Qualitative Materials 48 Implementation plan and variation 48 Participants 49 Key Informant Interviews 49 TOBACCO TAXATION SIMULATION MODELLING (TAXSIM) 50 KEY LIMITATIONS 50 Recall bias 50 Measuring FGD members’ socio-economic status 51 FINDINGS 52 TOBACCO 53 Tonga’s tobacco tax structure 53 Price elasticity 53 FOOD 62 Mutton flaps 66 Turkey tails 68 Chicken leg quarters 70 Ice cream 71 Imported instant noodles 73 Sugar-sweetened beverages 74 4 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga ALCOHOL 76 Beer 76 Spirits 79 EFFECTS OF CONSUMPTION TAX EXEMPTION ON IMPORTED HEALTHY PRODUCTS 81 Apples 82 Oranges 83 VIEWS FROM COMMUNITIES 84 FOOD 85 TOBACCO 90 ALCOHOL 91 FUTURE OF NCD TAX 92 DISCUSSION 94 HOW MUCH DID CONSUMERS RESPOND TO TAX-INDUCED PRICE CHANGES? 98 WHAT ARE THE KEY SUBSTITUTES TO TAXED PRODUCTS? 100 ILLICIT TRADE 101 HOW NCD TAX REVENUE IS USED 102 CONCLUSIONS AND RECOMMENDATIONS 104 CONCLUSIONS 105 RECOMMENDATIONS 106 Tax structure 106 Monitoring NCD tax implementation 106 Transparency in use of NCD tax 108 Complementary (non-tax) measures 108 BIBLIOGRAPHY 111 ANNEXES 114 ANNEX 1 115 ANNEX 2 119 ANNEX 3 123 ANNEX 4 125 ANNEX 5 127 5 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga List of Figures Figure 1: Map of Tonga 24 Figure 2: Type 2 diabetes prevalence trend among adults ages 25–64 years, Tonga, 1978–2012 25 Figure 3: GDP and health expenditure per capita, 2009–2014 26 Figure 4: Prevalence of tobacco consumption in the Pacific in 2016 29 Figure 5: Expenditure on unhealthy food, imported food, processed food, Pacific Island Countries (selected), 2014 31 Figure 6: Percentage of the population with overweight and obesity (≥25 kg/m²) in Pacific Island Countries 34 Figure 7: Tax per unit and retail price of imported cigarettes, T$ 54 Figure 8: Tax per unit and retail price of locally manufactured cigarettes, T$ 56 Figure 9: Tax revenue, all cigarettes, T$ 56 Figure 10: Sales volume, all cigarettes, T$ 56 Figure 11: Volume and tax revenue of imported cigarettes 57 Figure 12: Production volume and tax revenue of locally manufactured cigarettes 57 Figure 13: Cigarette affordability, 2013–2017 57 Figure 14: Market share of different cigarette brands in Tonga, 2011–2017 58 Figure 15: Reasons for behavior change 58 Figure 16: Behavior change following FY2016–7 (baseline) tobacco tax increase and FY2017–8 (endline) tobacco tax increase 59 Figure 17: Behavioral responses on cigarette consumption by wealth quintiles after the FY 2016/7 increase in excise tax 60 Figure 18: Proportion of manufactured cigarette smokers 60 Figure 19: Proportion of Tapaka Tonga smokers 60 Figure 20: Retail prices of Tapaka Tonga and manufactured cigarettes over time 61 Figure 21: Spending on manufactured cigarettes in the past week, T$ 61 Figure 22: Spending on Tapaka Tonga, T$ 61 Figure 23: Type of smokers in Tonga (November 2017) 62 Figure 24: Price rises after the excise tax/import duty increase 64 Figure 25: Awareness of the excise tax increase in July 2016 on selected food products as of June 2017 65 Figure 26: Knowledge about food products subjected to tax increase in July 2016 65 Figure 27: Notices about price changes of selected products by wealth quintiles 65 Figure 28: Tax per unit and retail price of mutton flaps 66 Figure 29: Consumption, tax revenue and total tax per unit of mutton flaps 67 Figure 30: Change in consumption of mutton flaps (baseline) 67 Figure 31: Change in consumption of mutton flaps (endline) 67 Figure 32: Consumption behavior change following the imposition of tax on mutton flaps 67 Figure 33: Reason for changing consumption behaviors of mutton flaps 68 Figure 34: Frequency of consumption of mutton flaps (baseline and endline surveys) 68 Figure 35: Tax per unit and retail price of turkey tails 69 Figure 36: Consumption, tax revenue and total tax per unit of turkey tails 69 Figure 37: Changed consumption behaviors of turkey tails at baseline 69 Figure 38: Changed consumption behaviors of turkey tails at endline 69 Figure 39: Behaviors following the imposition of tax on turkey tails 70 6 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 40: Reasons for changing consumption behaviors of turkey tails 70 Figure 41: Tax per unit and retail price of chicken legs 70 Figure 42: Consumption, tax revenue and total tax per unit of chicken legs 71 Figure 43: Changed consumption behaviors of chicken leg quarters at baseline 71 Figure 44: Frequency of consumption of chicken leg quarters (baseline and end line surveys) 71 Figure 45: Tax per unit and retail price of ice cream 72 Figure 46: Consumption, tax revenue and total tax per unit of Ice cream 72 Figure 47: Changed consumption behaviors of ice cream at baseline 72 Figure 48: Changed consumption behaviors of ice cream at endline 72 Figure 49: Frequency of consumption of ice cream (baseline and endline surveys) 72 Figure 50: Tax per unit and retail price of imported instant noodles 73 Figure 51: Consumption, tax revenue and total tax per unit of imported instant noodles 73 Figure 52: Notice of price change of imported instant noodles at baseline 73 Figure 53: Changed consumption behaviors of imported instant noodles at baseline 73 Figure 54: Tax per unit and retail price of sugar-sweetened beverages 74 Figure 55: Consumption, tax revenue and total tax per unit of SSBs 75 Figure 56: Notice of price change of SSBs at baseline 75 Figure 57: Changed consumption behaviors of SSBs at baseline 75 Figure 58: Frequency of consumption of SSBs (baseline and endline surveys) 75 Figure 59: Tax per unit of imported beer 77 Figure 60: Tax per unit of local beer 77 Figure 61: Tax per unit of beer 77 Figure 62: Consumption, tax revenue and total tax per unit of imported beer 77 Figure 63: Consumption and tax revenue, locally manufactured beer 78 Figure 64: Consumption, tax revenue and total tax per unit of beer 78 Figure 65: Changes in beer consumption frequency 78 Figure 66: Weekly beer consumption 78 Figure 67: Change in beer consumption behaviors 79 Figure 68: Reasons for changing alcohol consumption behaviors 79 Figure 69: Tax per unit of imported spirits 79 Figure 70: Tax per unit of locally manufactured spirits 80 Figure 71: Tax per unit of imported spirits 80 Figure 72: Consumption, tax revenue and total tax per unit of imported spirits 80 Figure 73: Consumption, tax revenue and total tax per unit of locally manufactured spirit 80 Figure 74: Consumption, tax revenue and total tax per unit of spirits 81 Figure 75: Consumption frequency of spirits 81 Figure 76: Public knowledge of food products subject to tax exemption, July 2016 81 Figure 77: Change in price after the tax exemption policy 82 Figure 78: Quantity of apples following the tax exemption policy 82 Figure 79: Apple consumption behaviors 82 Figure 80: Elasticity of taxed products 98 Figure 81: Elasticity of cigarettes 99 Figure 82: Annual budget of the Government of Tonga (GOT) and Ministry of Health 103 Figure 83: Distribution of households by island 123 7 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga List of Tables Table 1: Factors influencing health status for males and females, STEPS surveys 2004 and 2012 27 Table 2: Prevalence of tobacco use, 2004 and 2012 29 Table 3: Share of food expenditure in selected Pacific Island Countries, 2005–2010 32 Table 4: Fruit and vegetable consumption in Tonga, 2004 and 2012 32 Table 5: Physical activity measurements, 2004 and 2012 33 Table 6: Body mass index measurements in Tonga, 2012 35 Table 7: Total sample size of baseline data 42 Table 8: Total sample size of endline data 43 Table 9: Descriptive statistics of baseline and endline data 45 Table 10: Focus group discussion categories 47 Table 11: Tonga’s tobacco tax structure 53 Table 12: Price and total tax of a pack of 20 cigarettes (most sold brand) as percentage of retail price, Tonga (compared with selected Pacific countries), 2016 55 Table 13: Tax structure of food and SSB (July 2014–June 2018) 63 Table 14: Tonga alcohol tax structure 76 Table 15: Observed and desired behavior change and dietary outcomes 96 Table 16: Key substitutes of taxed food products 100 Table 17: Annual budget of the Government of Tonga (GOT) 103 Table 18: Total sample size of baseline data 115 Table 19: Total sample size of endline data 116 Table 20: Descriptive statistics of baseline data 117 Table 21: Baseline descriptive statistics by island 118 Table 22: Descriptive statistics of endline data 119 Table 23: Endline descriptive statistics by island 121 Table 24: Head of household distribution by gender 123 Table 25: Comparison of unweighted and weighted results 123 Table 26: Gender distribution 124 Table 27: Calculation of SSB elasticity 126 8 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga List of Images Picture 1: Tongan lady with her home vegetable garden right in the middle of Nuku’alofa (capital of Tonga) 21 Picture 2: The LDS Church introduces a “keyhole” technique for Tongan communities with limited space to grow vegetables 21 Picture 3: Packs of Tapaka Tonga as sold in shops 61 Picture 4: Pre-cooked mutton flaps 66 Picture 5: Cooked mutton flaps 66 Picture 6: Pre-cooked turkey tails 68 Picture 7: Cooked turkey tails 68 Picture 8: Apples sold above maximum price set by the Competent Authorities 83 Picture 9: Oranges as commonly sold in Tonga – in bags and not weighed 83 Picture 10: Oranges sold above the maximum price set by the Competent Authorities Picture 11 : “Intercrop” home garden right behind the Central Bank 110 of Tonga ― lettuce, onions, and kava Photo Credit Sutayut Osornprasop, Mona Palu, Sione Hufanga, Pita Feimoefiafi, Catherine Latu, Kevalin Saksiamkul, Don Mammoser, Shutterstock.com, Pinkgator Co. Ltd. List of Boxes Box 1: Government of Tonga tobacco control activities: a snapshot 37 Box 2: The impact of food tax on consumption among different socio-economic groups 86 Box 3: Response to the food tax – urban and rural differences 89 Box 4: Impact of taxes on retailers 92 Box 5: Women’s and men’s perspectives on NCD tax 93 Currency Equivalents Exchange rate effective as of 2017 Currency unit = T$ (Tonga Pa’anga) T$2.21 = US$1.00 9 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Abbreviations BMI Body Mass Index CEO Chief Executive Officer CIF Cost, Insurance, and Freight COPD Chronic Obstructive Pulmonary Disease CNMI The Commonwealth of the Northern Mariana Islands CVD Cardiovascular disease DM Diabetes Mellitus DOS Department of Statistics FAO Food and Agriculture Organization of the United Nations FGD Focus Group Discussion FSM The Federated States of Micronesia FY Fiscal Year GDP Gross Domestic Product GNI Gross National Income GOT Government of Tonga GPS Global Positioning System HIES Household Income and Expenditure Data IPC Income Purchasing Capacity LDS The Church of Jesus Christ of Latter-day Saints MCCTIL Ministry of Commerce, Consumer, Trader, Innovation and Labour MF Ministry of Fisheries MAFF Ministry of Agriculture, Forestry and Food MET Metabolic Equivalent of Task MFNP Ministry of Finance and National Planning MOH Ministry of Health NCD Noncommunicable Disease NHERC National Health Ethics and Research Committee NNCDC The National NCD Committee OECD Organization for Economic Cooperation and Development PCA Principle Component Analysis PCAE Per Capita Calorie Intake Per Day PIC Pacific Island Country PNG Papua New Guinea PPS Probability Proportional to Size RIP Relative Income Price SDA Seventh-day Adventist Church SI Socio-economic Index SMA Special Management Area Program SPC The Pacific Community SSB Sugar-sweetened Beverage STEPS STEPwise approach to surveillance TaXSiM Tobacco Tax Simulation TCC Tonga Communications Corporation TSDF Tonga Strategic Development Framework UN United Nations WB World Bank WHO World Health Organization WPRO Western Pacific Region Office 10 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Glossary BMI is an index of weight-for-height commonly used to classify overweight and obesity on a large population basis, measured in a Body mass index person’s weight in kilograms divided by the square of his/her height in (BMI) meters (kg/m²). In adults, overweight is defined as a BMI of 25 or more, whereas obesity is a BMI of 30 or more. In general, consumption tax is a tax on the purchase of a good or service. Consumption taxes can take the form of sales taxes/value added Consumption tax tax, tariffs, excise, and other taxes on consumed goods and services. However, Consumption Tax applied in Tonga is equivalent to value added tax instead of the broad definition. Customs duty or Customs duty, or import duty, is a tax paid for goods that are transported import duty across international borders. An excise tax is an indirect tax on the sale of a particular good or service such as fuel, tobacco and alcohol. Indirect means the tax is not directly paid by an individual consumer. Instead, it is levied on the producer or Excise tax merchant, who passes it onto the consumer by including it in the product’s price. Excise levied on cigarettes, alcohol and selected unhealthy products are sometimes called sin taxes because of high social cost of consuming this kind of goods. MET is the objective measure of the ratio of the rate at which a person expends energy, relative to the mass of that person, while performing some specific physical activity compared to a reference, set by Metabolic convention at 3.5 ml of oxygen per kilogram per minute, which is roughly Equivalent of Task equivalent to the energy expended when sitting quietly (equivalent to a (MET) caloric consumption of 1kcal/kg/hour). It is estimated that compared with sitting quietly, a person's caloric consumption is three to six times higher when being moderately active (3 ― 6 METs) and more than six times higher when being vigorously active (>6 METs) Noncommunicable diseases (NCDs) in a broader sense means diseases of non-transmissible/infectious nature. They are more commonly referred to as chronic diseases that tend to be of long duration and are the result of a combination of genetic, physiological, environmental, and behavioral factors. The four main types of NCDs are cardiovascular disease, diabetes, Noncommunicable cancer, and chronic lung disease; together they are collectively responsible diseases (NCDs) for almost 70 percent of all deaths worldwide. We refer to NCDs related to diet and nutrition as diet-related NCDs, which includes cardiovascular disease (such as heart attacks and stroke, and often linked to high blood pressure), diabetes, and certain cancers. Unhealthy diets and poor nutrition are among the top risk factors for these diseases globally. 11 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Glossary NCD-related taxes are implemented as a way to use fiscal policies, particularly taxation, to alter retail prices in such a way that sales and consumption of tobacco, alcohol and foods that are associated with non- communicable diseases. For example, taxes are levied on tobacco, NCD-related taxes alcohol, and foods high in specific nutrients/ingredients (salt/fats/sugar) or otherwise classified as “unhealthy,” including sugar-sweetened beverages. It could also take a form of reduced taxes (or subsidies) to promote the increased consumption of healthier food items, such as fruits and vegetables. A sin tax is an informal term used to describe an excise or ad valorem tax specifically levied on certain goods deemed harmful to society and individuals — for example, alcohol, tobacco, sugar-sweetened/soda Sin taxes drinks, unhealthy food, and gambling, among others — to increase their price in an effort to lower their use and reduce the negative impacts of the taxed substance. Sugar-sweetened beverages (SSBs) are any liquids that are sweetened with added sugars (for example, brown sugar, corn sweetener, corn syrup, dextrose, fructose, glucose, high-fructose corn syrup, honey, lactose, malt syrup, maltose, molasses, raw sugar, and sucrose), such as regular non-sugar-free soda, fruit/sports/energy drinks, sweetened waters, and coffee/tea beverages with added sugars. The calories in Sugar-sweetened SSBs can contribute to weight gain and provide little to no nutritional beverages (SSBs) value, and lead to other health risks including obesity, tooth decay, heart disease, and type 2 diabetes. SSBs are associated with minimal effect on hunger, so food intake does not decline when consuming sugary beverages. With an aim of reducing the consumption of SSBs, the sales of the defined products are regulated — for example, they are restricted in school premises, or taxed as an important public health measure. 12 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Acknowledgments This report was prepared by Sutayut Ministry of Agriculture, Forestry and Osornprasop (Task Team Leader), Rong Food (MAFF), Ministry of Fisheries (MF), Zheng, Sione Hufanga, Catherine Latu, Ministry of Commerce, Consumer, Trade, Kiatanantha Lounkaew, Jutta Krahn, Innovation & Labour (MCCTIL), as well as Erling Ve’etutu, Sela Latailakepa, and the Department of Statistics (DOS) and Sarulchana Viriyataveekul. Yang Wang, Tonga Health Promotion Foundation. In Xiao Hu, Sione Saumaki, Kilifi Lolohea, addition, the team highly appreciates the and Pita Feimoefiafi provided research valuable inputs from the National Food support. The study was conducted under Council, church leaders, Tonga Family the guidance of Enis Baris, Toomas Palu, Health Association, Civil Society Forum Michel Kerf, and Lasse Melgaard. The of Tonga, and the Tonga Trust, as well as team sincerely appreciates the excellent representatives from international communication support from Jodie Brabin, organizations, development partners, and valuable in-country liaison support from the private sector. We sincerely appreciate Tatafu Moeaki and Natalia Latu, as well the people of Tongatapu and Vava’u who as superb administrative and logistics agreed to be part of the surveys and focus support from Pimon Iamsripong, Elizabeth group discussions, as well as all the stake Ika, Roberto Belisario, Abigail Blenkin, holders from the government, private Farazeen Amjad, and Kevalin Saksiamkul. sector, church, civil society organizations, The research team sincerely appreciates and development partners who participated the overall support from the following and contributed to the stakeholder individuals from the World Bank: Susan consultation meetings in Tongatapu and Ivatts, Patricio Marquez, Meera Sheekar, Vava’u. Maude Ruest, Aparnaa Somanathan, Venkatesh Sundararaman, Kim Edwards, The team highly appreciates the valuable Robert Utz, Lori Geurts, Sheila Dutta, support from the following individuals Ashi Kathuria, Phoebe Folger, and Jesse Doyle. The team is grateful for the valuable Government of Tonga comments provided by the following Ministry of Agriculture, Forestry and Food: technical peer reviewers from the World Hon. Mrs. Losaline Ma’asi, Dr. Viliami Bank: Maria Eugenia Bonilla-Chacin (Lead Manu (CEO), Ms. Mele ‘Atuekaho, Mr. Economist), Robert Johann Utz (Program Metuisela Falesiva, Ms. Leody Vanikolo, Leader and Lead Economist), Caryn Ms. Kakau Leander Bredenkamp (Senior Health Economist), and Alan Fuchs (Senior Poverty Economist). Ministry of Commerce, Consumer, Trader, Innovation & Labour: The team sincerely appreciates the excellent Mr. Edgar Cocker (CEO), Mrs. Sandra collaboration with the Government of Fifita, Mr. Anthony Cocker Tonga, which contributed to the completion of this study. The team highly values Ministry of Education and Training: the support provided by the Ministry Hon. Mr. Penisimani Fifita, Mr. Sione of Finance and National Planning Tokoma’ata (MFNP), Ministry of Revenue and Customs (MORC), Ministry of Health Ministry of Finance and National Planning: (MOH), Ministry of Internal Affairs (MIA), Hon. Rev. Dr. Pohiva Tu’I’onetoa (Minister), Mr. Tevita Lavemaau (Former Minister), 13 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Ms. Balwyn Fa’otusia (CEO), Ms. Tonga Department of Statistics: Kilisitina Tuaimei’api, Mr. Sione Dr. Viliami Fifita (Government Statistician), Faleafa, Ms. ‘Ana Lotuma Falefehi Mr. Sione Lolohea, Mr. Winston Ika, Ms. Elizabeth Baker, Ms. Kilisitina Fainga’anuku Angahiki Puleiku, Mr. Semisi Faupula Tonga Health Promotion Foundation: Ministry of Fisheries: Ms. ‘Ofeina Filimoehala (CEO), Ms. Seini Hon. Mr. Semisi Fakahau Filiai (former CEO), Ms. Jeanie McKenzie, Mrs. Monica Tu’ipulotu Ministry of Health: Hon. Dr. Saia Piukala (Minister), Dr. Siale Church leaders ‘Akau’ola (CEO), Mr. Semisi Fukofuka, Dr. Pastor Saia Vea (President of the ‘Ofa Tukia, Mr. Sioape Kupu, Mr. Eva Mafi, Seventh Adventist Church), Dr. Tuípulotu Ms. Tinalasa Vunipola, Ms. Latu Fusimalohi, Katoanga (President of the Church Mr. Tupou Tui’ilautala, Dr. John Lee Taione of Tonga), Mr. Silongoatonga Samani (Head of Welfare and Humanitarian of Ministry of Infrastructure: the Church of Jesus Christ of Latter Mr. Petelo Sanele Ulupano -day Saints), Rev. Dr. Mele’ana Puloka (President of Free Wesleyan Church Ministry of Internal Affairs: Education), Rev. ‘Etuate Manuofetoa Hon. Dr. Saia Piukala (Minister), Mr. (Free Wesleyan Church), Father Pauli ‘Onetoto ‘Anisi (Acting CEO), Ms. Ana Ta’ai (Catholic Church), Mr. Viliami Bing Fonua (former CEO), Ms. ‘Eva Talivakaola (Church of Jesus Christ Tuipeatau - Tu’uholoaki, Mr. ‘Alotaisi of Latter-day Saints) Takau, Mr. Penisimani Vainikolo, Ms. Malianive Finau, Mr. Kotoni Fifita, Mr. Development partners Mapa Hafoka, Mr. Ula Kaufusi, Mr. Vava Food and Agriculture Organization of the Lapota United Nations: Eriko Hibi, Ramani Wijesinha-Bettoni, Ministry of Police: Fiasili Lam, Pau Molevuka Likiliki Mr. Taulango Tapueluelu The Government of Australia’s Depart Ministry of Revenue and Customs: ment of Foreign Affairs and Trade: Hon. Mr. Mateni Tapueluelu (Minister), H.E. Andrew Ford, Rebecca Dodd, Mr. Kelemete Vahe (CEO), Mr. Kulu ‘Anisi Madeleine Scott, Chris Sturrock, Kirsty Bloomfield (former CEO), Mr. Heiloni Dudgeon, Jane Pepperall, Debra Allan Latu, Mr. Vehitau Ma’ukie, Mr. Tupou Latavao, Mrs. Melenaite Ma’ukie The Government of New Zealand’s Ministry of Foreign Affairs and Trade: Ministry of Tourism: H.E. Tiffany Babington, Sumi Subramaniam, Mr. Bruno Toke Elena Procuta, Olivia Fukofuka, Anna Pasikale, Sarah Johal National Reserve Bank of Tonga: Ms. ‘Ana Soakai Takai The Government of Japan: H.E. Tetsuya Ishii, Yasushi Hayashi, Saito Parliament of Tonga: Sho Hon. Siaosi Sovaleni (Member of Parliament) World Health Organization: Yutaro Setoya, Wendy Snowdon Vava’u Governor’s Office: Hon. Lord Fakatulolo 14 The Pacific Community (SPC): Paula Vivili, Sunia Soakai Academics University of the South Pacific, University of International Business and Economics (China) Civil Society Organizations Mrs. Mele Amanaki, Mrs. Papiloa Foliaki, Mrs. Amelia T. Hoponoa, Mr. Tevita Halahala, Ms. Fuiva Kavaliku, Mr. Sione Taumoefolau, Mr. Rhema Misa, Mr. Sione Fifita, Pastor Havili Sefesi, Mrs. Lisa Fanua, Mr. Siaosi Moala Private Sector Mrs. Maliana Tohi, Mr. Min Seng Cai, Ms. Elizabeth Cowley The World Bank team very much appreciates the close collaboration with Ms. Ramani Wijesinha-Bettoni and team from FAO during the conceptualization and household questionnaire design stages, particularly in working together to publish Effects of Food Taxation in Tonga: A Snapshot report (http://www.fao.org/3/a- i8052e.pdf). The preliminary findings from the Snapshot have also been confirmed by this report. The team is grateful for the funding support from the World Bank’s Tobacco Control Program, the Government of Australia (Department of Foreign Affairs and Trade), the Government of New Zealand (Ministry of Foreign Affairs and Trade), and the Government of Japan through the Japan Trust Fund for Scaling Up Nutrition. Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Executive Summary 16 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Tonga has experienced a high burden This study aims to generate policy- of noncommunicable diseases (NCDs) relevant findings and insights on the in the past two decades. Data indicate implementation of taxation policy on the continuing rise of four major NCDs – tobacco, alcohol, and unhealthy food cardiovascular disease, diabetes, cancer, consumption behaviors; as well as and respiratory diseases – as the leading pricing, government revenues, and the causes of premature death and disability. country’s response to Tonga’s rising According to latest data, NCDs accounted NCD burden. The findings are expected for four out of five leading causes of to help strengthen the monitoring and mortality in Tonga, and 99.9 percent of evaluation of NCD-related fiscal policies, Tongan adults aged 25–64 are at moderate not only of behavior change but also revenue to high risk of developing an NCD. generation. The findings also intend to help improve the design of future The increase in behavioral-related risk NCD-related fiscal policy – as well as factors such as smoking, poor diet other, non-fiscal policies – to address (e.g. high in sugar, salt, trans-fat and NCD challenges. These lessons from saturated fat, and low in fruit, vegetables, Tonga will also be relevant and beneficial legumes, fish etc.), harmful alcohol to other Pacific Island Countries (PICs) intake, and physical inactivity are that face similar NCD and obesity burdens. acknowledged as the major contributing factors to the rise in NCDs in Tonga. This study employs the following According to its 2012 STEPS Survey, assessment tools and sources of Tonga has one of the highest rates of data: (i) government administrative data; overweight and obesity (≥25 kg/m2) in (ii) baseline and endline household the world – reaching 90.7 percent among surveys, focusing on the impact of NCD tax some population groups. Both obesity on consumption behaviors in Tongatapu and overweight are common at younger and Vava’u which account for over 85 ages. Furthermore, almost one in two percent of Tonga’s population; (iii) retail men smoke, and smoking appears to be surveys focusing on the impact of the increasing among young women. NCD tax on prices; (iv) tobacco taxation simulation modeling (TaXSiM); (v) a In response to the rise in NCDs, in recent qualitative study that includes focus group years the Government of Tonga has discussions and in-depth interviews. gradually introduced taxation measures to influence consumption behaviors Positive impacts from the NCD tax policy of the population, complementing have been observed. The excise tax existing regulatory and legislative resulted in a price increase across all interventions. The increase in excise tax the taxed products. The excise tax made on imported manufactured cigarettes in cigarettes less affordable and affected July 2016 was of particular significance. smokers’ behaviors. Tax on cigarettes Unlike the small, incremental increases has a greater effect on “less well-off” of previous years, the excise tax in July smokers, as a larger number of less 2016 was raised by nearly 50 percent well-off smokers reduced consumption to T$380/1,000 sticks, compared with of manufactured cigarettes. This is T$255/1,000 sticks in the previous year. At consistent with the 2018 Lancet findings the same time, the government waived on the equity impacts of price policies. The consumption tax on imported fruits, in study found that low-income households anticipation that the price would fall, and that respond to price changes more readily than the population would in turn increase its consumption of these more healthy products. 17 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga higher income households. Furthermore, Tapaka Tonga – the main unhealthy price policies affect the spending of a larger product that smokers use as a substitute number of high-income households than for manufactured cigarettes – is low-income households, and any resulting affordable and widely available in price increase tends to be financed shops. According to the Ministry of disproportionately by high-income Revenue and Customs, over 90 percent households. Price, rather than health, is 1 of Tongan shops sell Tapaka Tonga. the main reason among those who decided According to interviews, retailers buy to change behaviors and reduce Tapaka Tonga from manufacturers at T$2- consumption. 3 per pack, and retail it to consumers at T$6-7. Tapaka Tonga is currently not taxed Nevertheless, the positive impacts of the and is a very profitable product. tobacco tax policy in Tonga are diluted. There was a shift from more expensive Increased excise tax has led to brands of cigarettes to less expensive decreases in the frequency and brands, particularly locally manufactured amount of alcohol products consumed; cigarettes, which are subject to less tax. however, a shift to substitutes But more importantly, there was a big has been observed. Significant numbers shift from manufactured cigarettes to the of drinkers have shifted to kava. There untaxed local hand-rolled tobacco leaves is also increased incidence of substance called Tapaka Tonga. Today, most tobacco abuse, including marijuana, glue/petrol users in Tonga smoke Tapaka Tonga. sniffing, and crystal methamphetamine, though it is not clear if this is related to the Tapaka Tonga is no less harmful to excise tax rise. health than manufactured cigarettes. Many Tongans wrongly believed Tapaka Effects of excise tax on food are mixed. Tonga to be less harmful than Excise tax on turkey tails, mutton flaps, and manufactured cigarettes, and some ice cream helped reduce consumption even thought it was “healthy” given of these products – though the tax had that it has been marketed as “organic very limited effects on the consumption tobacco”. However, according to a of chicken leg quarters. In addition, the World Health Organization analysis, tax on chicken leg quarters appeared to like manufactured cigarettes, Tapaka have regressive effects on less well-off Tonga contains many toxic chemicals, households. Despite the tax-induced price, including, nicotine, formaldehyde, chicken leg quarters remained the cheapest acetaldehyde, acetone, acrolein, propinal meat in the market, and less well-off dehyde, crotonaldehyde, butryraldehyde, households continue to buy it. Hence, and 2-butanon. The amount of nicotine in the tax on chicken leg quarters took a Tapaka Tonga is proportionately higher proportionately greater amount of financial than that of manufactured cigarettes.2 resources from those on lower incomes. 1 Sassi, F., et al, Equity impacts of price policies to promote healthy behaviours, The Lancet Taskforce on NCDs and economics, Volume 391, Issue 10134, P2059-2070, May 19, 2018. 2 Health Science Authority of Singapore report, 2 April 2013 (obtained from World Health Organization). 18 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The potential risks of swapping to The exemption of consumption tax other, cheaper, and/or low quality on selected healthy products (e.g., unhealthy food products are substantial, imported fruits) resulted in neither given the variety of food choices in lower retail prices nor higher consumption Tonga. While unhealthy food products of these products. Imported fruit prices which are subject to excise tax have remained the same, despite the become more expensive, there are still removal of the 15 percent consumption unhealthy food products which are not tax. The benefits from the tax exemption subject to excise tax, making them more went to traders rather than consumers.3 affordable than the taxed products. Forty percent of mutton flap consumers reported Tax-induced price changes have caused shifting to substitutes – salted beef, tinned Tongan consumers to reduce their fish and corned beef. Salted beef and consumption of unhealthy food corned beef were not subject to excise tax, products, albeit to different degrees. but they also contain high and unhealthy Instant noodles and chicken legs have price levels of salt and fat. The main substitutes elasticity of -0.38 and -0.39 respectively. of ice cream and instant noodles were The negative sign indicates that price identified as locally manufactured ice and consumption vary inversely. Thus, for cream and instant noodles, which were instant noodles, a 1 percent increase in not subject to excise tax at the time of the price will reduce instant noodle consumption survey, making them cheaper than their by 0.38 percent; similarly, a 1 percent imported counterparts. This demonstrates increase in the price of chicken legs will the importance of imposing unified tax reduce consumption of chicken legs by 0.39 rates for both imported and locally percent, which shows that consumption manufactured unhealthy products. of these products did not change Otherwise, people could easily shift to significantly as compared to changes consuming locally manufactured unhealthy in price. Elasticity of sugar-sweetened products, defeating the health outcome beverages is -0.93 which is within the purpose of the tax. ranges of WHO’s findings, and is more sensitive to price change compared The availability and affordability of to the previous two products. The two healthy food alternatives remain a products most responsive to price change major issue in Tonga. Therefore, are ice cream and mutton flaps with more policy interventions are needed elasticity of -1.38 and -1.92 respectively. to make healthy food more affordable and accessible. For example, fresh Smokers from lower wealth quintiles fish is highly desirable, but less well-off are more responsive to tax-induced households reported that they cannot price changes and are likely to reduce consume fresh fish on a regular basis due cigarette consumption as the price of to the high price. Households in general cigarettes increases. Overall elasticities are hard-pressed to consume healthy of demand for cigarettes, following the alternatives because of the lack of July 2017 tax increase, was -0.39; a price availability, accessibility, and high cost. increase of 1 percent will lower cigarette demand by 0.39 percent. However, there was great variation of behaviors of 3 While Tonga has a price control mechanism for imported fruits, it seems the policy has not been effectively enforced. 19 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga smokers in different socio-economic fact that, during the period of this study, groups. The elasticities of demand for there was no other variable that caused cigarettes among smokers from the lowest significant change to prices on these wealth quintile was -0.63, compared with products also lends support to the claim -0.21 from the highest wealth quintile. that tax imposition was the major driver for This means that smokers in lower wealth lower consumption of the taxed products. quintiles were more responsive to price change. Recommendations To conclude, how much a consumer Tax structure responded to price change depended on the characteristics of the product. Commercialized Tapaka Tonga should For a product with a low price per unit, a be subject to excise tax, as it is harmful change in price would not have a significant to health yet cheap and readily impact on total consumer and household available in shops. The fact that Tapaka expenditure; consumption of these Tonga is not subject to excise tax also products, therefore, would be less sensitive contributes to the reduction in government to price change. Products whose unit revenues from excise tax on tobacco price occupied a noticeable proportion products, as many smokers have substituted of household expenditure would be more the more expensive taxed manufactured sensitive to price change. Elasticity also cigarettes with untaxed Tapaka Tonga. depended on the nature of the product itself. Products which were considered Excise tax on chicken leg quarters should a daily staple would be less sensitive to be removed for the following reasons: price change. Luxury products would be • Lack of global evidence that chicken legs more sensitive to price change. Lastly, are “unhealthy”, as their nutritional benefits addictive products such as cigarettes depend on how they are cooked (e.g. would be less sensitive to change in price. remove fatty parts and skin, not deep-fried). When wealth quintile is taken into account, • Chicken leg quarters are an important smokers from lower wealth quintiles source of protein for Tongan children and are more likely to reduce their adults, and despite the tax-induced price hike consumption behaviors as price increases. they remain the cheapest meat option on These results provide empirical support for the market. the effectiveness of tax as an instrument • Most households continue to buy chicken to divert Tongan consumers from consuming leg quarters despite the price increase, products that are detrimental to their long- creating a regressive effect on poorer term health and well-being. households. While it cannot be concluded that tax The government should take a consistent increases were the sole reason for approach and apply scientifically reduction in consumption of taxed supported criteria across all food products, the tax can be said to have groups, including the use of nutrient had significant influence on the profiling that indicates unhealthy consumption pattern of Tongan threshold (e.g. levels of fat, salt, and consumers. For example, around 76 to 79 sugar), in designing the NCD tax percent of smokers said that they smoked policy. This should be done in close less because of the price change. Similarly, consultation with health and nutrition around 98 to 99 percent of mutton flap experts from the Ministry of Health, Ministry consumers reduced their consumption of Agriculture, Forestry and Food, Tonga because of the change in price. The Health Promotion Foundation, the National 20 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Food Council, and relevant development to assess if the price per unit has decreased partners. This should be applied to or increased. Use of weighing scales for both food products subject to an increase selling fruits and vegetables should be in excise tax, as well as food products enforced, and it will not only help the subject to consumption tax exemption. government in monitoring prices, but also empower consumers in ensuring that traders NCD-related unified excise tax should do not overcharge. The empowerment of be applied to both imported and locally consumers in this way is essential, produced products (and at the same given that the government may not have level), or its effects will be diluted. For sufficient human resources to conduct price example, excise tax on sugar-sweetened monitoring regularly throughout the country. beverages (SSBs) had earlier applied to only imported SSBs. Local businesses The government needs to monitor the were quick to note the loophole and started implementation of the NCD tax regularly to produce SSBs locally. Locally produced and make sure that annual tax increases SSBs were not subject to NCD tax well-exceed increases in consumer prices and have become cheaper than and incomes to ensure effectiveness. imported SSBs (locally produced SSBs Small, incremental increases in tax alone cost approximately T$1.5 while imported do not work. What drives behavior change SSBs cost around T$3.5 for a similar is the “price shock” that comes after a large type and size of SSB), leading to higher tax increase. After the large tax increase, consumption of locally produced SSBs. small incremental increases can be used This also applied to instant noodles, to sustain impact. But most importantly, the whereby imported instant noodles were government needs to monitor regularly and subject to excise taxes, while locally make sure that annual tax increases well- manufactured instant noodles were not, exceed the increase in consumer prices and have become cheaper than imported and incomes. Evaluation of interventions ones. In addition, the experience of imposing or programs needs to be planned and different excise tax rates on imported and designed before the start of interventions. locally manufactured cigarettes shows that Evaluation will generate evidence on it increases the price gap between imported impacts of the interventions and programs and locally manufactured cigarettes, leading as well as lessons learned. more smokers to switch to cheaper locally manufactured cigarettes. It is recommended There is room for improvement on that excise tax rates for imported and locally communication, through relevant media manufactured unhealthy products should channels, in particular mass media and be at the same level for the same products, social media as well as to organize public e.g. imported cigarettes and locally consultations to reach out to different manufactured cigarettes. stakeholders regarding the NCD-related taxation policy. From the household survey, Monitoring tax policy implementation there was a lack of knowledge among the majority of the population about the imposition Better monitoring of the prices of tax and increase in NCD-related taxation that exempt, healthy food is needed to became effective in July 2016, and again in July ensure that tax exemption achieves 2017. Many people still do not have a clear idea the goal of price reduction, and in of what unhealthy products are, what healthy turn higher consumption of healthy options are, and how to access them. food products. Most shops in Tonga do Consultations with key agencies, key stakeholders, not use weighing scales for fruits and church leaders, and representatives of the vegetables, making it difficult for consumers communities should be conducted prior to imposition of new taxation policy. 21 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Transparent use of NCD tax revenues initiative as an easy way to grow vegetables that can be afforded by every household, In response to calls from the even with limited land area. A keyhole garden communities, the Government of is a raised-bed system that requires minimal Tonga may explore how it can watering and fertilizer due to its ability to demonstrate to communities that retain moisture and use compost. Its unique the revenues acquired from NCD- design allows farmers to irrigate their crops related taxes are used to support with recycled water from household washing. and promote healthy lifestyles The keyhole structure also makes it easy among the population, as well as for farmers to add soil or manure to the improve health care services. Some of garden if crops need additional fertilizer. A the key health promotion interventions raised bed is also particularly effective for that communities desired from the those who have physical limitations, such government include: making healthy as the elderly or people with disability. food options cheaper, invigorating health There are also good examples of intercrop promotion activities, and supporting home vegetable gardens led by households, production of healthy food products, etc. even in the center of the capital Nuku’alofa, which provide good examples to Tongan communities. The government should Complementary (non-tax) measures also build on communities’ requests to support access to markets to stimulate More policy interventions are needed production, study experiences from poultry to make healthy food more accessible raising in other PICs (e.g. Fiji), and and attractive, and also to encourage increase community mobilization and consumption of smaller portions of organizations (e.g. farmers’ groups). food. For example, fresh fish is highly desirable, but less well-off households Making healthy food more affordable reported that they cannot consume fresh and accessible is not in itself adequate. fish on a regular basis due to the high It is important to introduce parallel price. The Ministry of Fisheries’ program interventions to make healthy food that required foreign fishing vessels to off- attractive, “yummy” and ensure livestock load at least 3 tons of fish to sell to the farms are sustainable. Making healthy food community at T$2.5-9 per kg has been very more affordable alone does not guarantee popular. However, it seemed to benefit only that the population will stop consuming communities living close to where the fish unhealthy food and swap to healthy were sold, e.g. Nuku’alofa. Communities products. Communities in focus group from other areas of Tongatapu and Vava’u discussions said they lacked knowledge of would like to benefit from a similar program. how to make vegetables more palatable In addition, the overall portion of food that at home, and requested cooking Tongans consumed each meal tend to demonstration programs to increase be large and at times excessive. Hence, the variety of dishes that could be eaten at consumption of smaller portions of food home. Education programs on how to should be encouraged and promoted. initiate and sustain chicken farms should also be introduced alongside livestock More studies should be conducted on programs to ensure long-term access to home production initiatives to generate healthy food. evidence and promote successful cases. For example, the government could build The increase in revenues from the on the successful Church of the Latter NCD tax, particularly from food, SSBs, Day Saints-supported “keyhole garden” and alcohol tax, has increased fiscal 22 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga space that will help the Government agriculture and fisheries, accompanied by of Tonga increase the resources to strong legislation, regulations, effective support health promotion/disease communication and campaigns, and an prevention activities as well as improve enabling environment. For example, the health care services to help the country government could complement taxation achieve universal health coverage. As intervention on Tapaka Tonga by regulating highlighted in the Pacific NCD Roadmap, a requirement for large and clear warning such complementary measures are essential labels on packets of Tapaka Tonga as well as to address the NCD crisis. This is even providing support or incentives (knowledge more important in light of the findings of and/or seeds) to encourage farmers to this study, which show that without switch from growing tobacco to growing complementary measures, tax increases healthy cash crops. may merely result in a switch from one form of unhealthy consumption to another. The National NCD Committee, its advisory There are lessons learned from several bodies, The National Food Council, and countries that use the revenue from sin Tonga Health Promotion Foundation taxes to support health promotion and should play an active role in promoting health services, which help these countries the multi-sectoral interventions to achieve universal health coverage. address the NCD crisis that are to be implemented by sectoral ministries. It is important to engage political, In the past two years the National NCD social and community leaders to Committee and the National Food Council support social and behavior change played a very limited role in the design, to address the NCD crisis. From the implementation, and monitoring of NCD qualitative study it is clear that political, tax and other health promotion activities. church, and community leaders are very The concerted efforts by these bodies, influential, and their roles in promoting combined with active implementation by healthy lifestyles are critical. There are relevant sectoral ministries, churches and superb experiences in healthy eating and civil societies can make a difference to healthy lifestyles that have been Tonga and help address the NCD crisis. implemented by selected churches and communities. Their lessons should be shared widely. Good examples include the introduction of healthy food at church conferences by the Seventh-day Adventist Church (SDA), which encouraged the use of healthy ingredients and discouraged overeating. The Church of Jesus Christ of Picture 1 : Tongan lady with her home vegetable garden Latter-day Saints (LDS) has also introduced right in the middle of Nuku’alofa (capital of Tonga) the “keyhole garden” initiative as an easy way to grow vegetables that can be afforded by every household, even with limited land area as shown in Pictures 1 and 2. Multi-sectoral interventions are needed to address NCD burdens; fiscal /taxation policy alone is not adequate. The fiscal/taxation policy needs to be complemented by other policy interventions Picture 2: The LDS Church introduces a “keyhole” technique for Tongan communities with limited space from sectors including health, education, to grow vegetables 23 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Introduction 24 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Tonga’s NCD burden is immense. Data of particular significance. Unlike the small, from the past two decades indicate the incremental increases of previous years, continuing rise of four major NCDs as the excise tax in July 2016 rose by nearly 50 leading causes of premature death and percent to T$380/1,000 sticks, compared disability in the country – cardiovascular to T$255/1,000 sticks in the previous disease, diabetes, cancer, and respiratory year. At the same time, the government diseases. According to the latest data,4 waived consumption tax on imported fruits NCDs accounted for four out of five leading (e.g. apples and oranges) in the hope causes of mortality in Tonga, and 99.9 that lower prices would encourage the percent of Tongan adults ages 25–64 are population to consume more fresh fruit. at moderate to high risk of developing an NCD. In response to the significant tax rise The increase in behavior-related risk in July 2016, the Government of Tonga factors such as smoking, unhealthy asked the World Bank to study the diets, harmful alcohol intake, and impacts of these fiscal interventions. physical inactivity are the major The aim is to understand whether Tongan contributing factors to Tonga’s rise consumers are responding to price rises in NCDs. Unhealthy dietary practices in resulting from higher taxation on tobacco, Tonga include high consumption of fats alcohol, and selected unhealthy foods and (including saturated fats from animal- beverages, and to lower-priced healthy based foods and trans-fatty acids from foods as a result of the waived consumption snacks and packaged foods), high intake of tax. The aim is also to understand how sugar and salt, as well as low consumption best to improve the use of these fiscal of fresh fruits, vegetables, and fresh fish. policies to address NCDs in the future. Tonga has one of the highest rates of overweight and obesity (≥25 kg/m²) in the This report presents findings and world, rising to as high as 90.7 percent recommendations from the World among Tongan adults. Both obesity and Bank’s Improving the use of fiscal overweight are common at younger ages. policy on tobacco, alcohol, food and Furthermore, almost half of all men smoke, beverages as a response to the NCD and smoking appears to be increasing crisis in the Pacific: A country case among young women. These are all study in Tonga. The study aims to strongly linked to “unhealthy environments”, generate policy-relevant insights into the and require policies, regulations and implementation of Tonga’s taxation policy legislation interventions, among others. in relation to tobacco, alcohol, food and beverages (hereafter referred to as “NCD In recent years the Government of tax policy”). These insights can enhance Tonga has gradually introduced taxation the discussion about the tax policy’s effect measures to influence the population’s on pricing, consumer behavior, and consumption behaviors in order government revenues, and generate a to address the rise in NCDs. The better understanding of communities’ increase in excise tax on imported viewpoints on the tax policy and measures manufactured cigarettes in July 2016 was other than fiscal interventions. 4 Ministry of Health, Kingdom of Tonga Risk Factors STEPS Report, 2014. 25 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The findings and recommendations Figure 1: Map of Tonga in this report are expected to help strengthen the monitoring and evaluation of NCD tax policy in relation to change in consumption behaviors and social norms, as well as in relation to broader government tax policy. The study also contributes to the future design of NCD tax and non-tax policies to address the NCD challenge. Key lessons from Tonga may also be relevant to other Pacific Island Countries (PICs) that face similar health challenges in helping them design evidence-based policies to address NCDs and obesity. The country context Tonga is a small Pacific island nation Source: World Atlas (accessed May 11, 2019), https://www.worldatlas.com. with a population of 106,000 (see Figure 1). It comprises over 170 islands and islets scattered over 800 km, of which around Similar to most other Pacific Island Countries 36 are inhabited. About 70 percent (PICs), smallness and remoteness pose of the population live on the main island challenges to Tonga, pushing up the of Tongatapu, where the capital, Nuku’alofa, cost of economic activity and limiting is located. Tonga is a constitutional monarchy both competitiveness in global markets that has recently undergone democratic and the scope for diversification. Average changes and is classified as an upper- economic growth over the past five years middle-income country with a gross has thus remained low by global standards, national income (GNI) per capita of averaging 1.2 percent annually, and has US$4,010 (US$6,050 PPP).5 been volatile.6 Despite being an upper-middle-income The country remains dependent on country, Tonga’s economy is fragile, external aid and remittances to offset a and its economic growth potential large trade deficit arising from a heavy is constrained by structurally high reliance on imported foodstuffs, fuels, costs of living, exposure to economic chemicals, machinery, and transport and environmental shocks, as well as equipment. There are limited opportunities natural disasters. Tonga is the world’s for formal, salaried employment in Tonga most geographically remote nation from and a large proportion of the population the major centers of economic activity. is engaged in subsistence agriculture and fishing. In recent years there has been significant migration of Chinese nationals 5 World Development Indicators, World Bank, 2018. 6 International Development Association program document for a proposed development policy grant to the Kingdom of Tonga for the Second Inclusive Growth Development Policy Operation, April 27, 2017. 26 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga to Tonga. Most have engaged in the with almost half (48 percent) of all deaths commercial sector. Most shops in Tonga occurring in the working-age population. today are owned and/or run by Chinese This is especially the case for men, who families. account for 68 percent of overall CVD deaths. The incidence of cancer has also The NCD crisis and development increased over time, with the high mortality impacts rate due partly to late detection.10 Traditionally Tonga has had one of the best overall levels of health of all PICs and still retains good maternal and Figure 2: Type 2 diabetes prevalence trend in Tongan adults aged 25―64 years, 1978―2012 child health outcomes. The country has seen a dramatic reduction in communicable diseases and maternal and child mortality since the 1950s. Relative to other small Pacific island states, Tonga has a sound primary health care system and public health infrastructure, and comprehensive antenatal and postnatal care, immunization, water, sanitation, and waste disposal programs.7 According to the Human Capital Index, 98 out of 100 children born in Tonga survive to age 5. Only eight out of 100 children in Tonga have stunted growth.8 While the total fertility rate remains high (3.6 in 2016), the maternal mortality Source: Lin, S. et al. 2016. Diabetes and obesity trends in rate is less than 40 per 100,000 live births.9 Tonga over 40 years. Asia Pacific Journal of Public Health Vol This, on average, means there are fewer 28(6): 475–485. than two maternal deaths per year. Nevertheless, Tonga has seen a sharp NCDs and their treatment costs rise in NCD prevalence, and NCDs now impose a significant burden not just account for four of the top five causes on individual patients but also on of death – cardiovascular disease (CVD), households, communities, employers, diabetes, cancer, and respiratory diseases. health care systems, and the country’s Accompanying this is a significant increase fiscal system. While GDP per capita tends in the number of premature deaths. For to rise and fall in Tonga, health expenditure example, the prevalence of diabetes per capita has seen a rising trend in recent has risen from less than 5 percent in years, with NCD treatment one of the 1970 to over 20 percent today (see major drivers (Figure 3). This presents Figure 2), with deaths from diabetes many financing challenges and cost slightly more common in females. CVD pressures to the country’s fiscal system. is a major cause of premature mortality, 7 Human Capital Index, World Bank, 2018. 8 The Kingdom of Tonga Health System Review, Health Systems in Transition, Vol. 5, No. 6, 2015. 9 Human Capital Index, World Bank, 2018. 10 The Kingdom of Tonga Health System Review, Health Systems in Transition, Vol. 5, No. 6, 2015. 27 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 3: GDP per capita and health expenditure NCD risk factors per capita, 2009―2014 In Tonga, four key NCD risk factors have been identified: tobacco consumption, unhealthy diet, lack of physical activity, and harmful alcohol use. A combination of unhealthy diet and lack of physical activity also contributes to the high obesity burden. Comparison of data between the 2004 STEPS survey11 and 2012 STEPS survey show alarming results, with poorer outcomes for several indicators (see Table 1). The percentage of adults with raised blood pressure rose from 23.1 per cent in 2004 to 27.6 percent in 2012. The percentage of adults with impaired fasting Source: World Bank, Pacific Possible, 2017. glycaemia rose from 17.1 percent in 2004 to 23.8 percent in 2012, while the percentage Tonga’s high NCD burden also affects of adults with raised fasting blood glucose labor productivity and economic growth. doubled over the same period – from NCDs affect not only the productivity 16.4 percent to 34.4 percent. It is of those they burden, but also that of important to note that these risks have risen healthy members of the workforce who markedly for women. By 2012, the percentage need to take leave to care for them. And in of Tongan women with impaired fasting terms of capital, the NCD burden not only glycaemia12 and raised blood pressure leads to the loss of income, savings, and had risen to a similar level as men, while investment at individual and household the number of women with raised fasting levels, but also imposes significant costs blood glucose had exceeded men by nearly to the government – diverting spending 10 percentage points. The percentage of and investments from areas beneficial Tongan women with raised total cholesterol to the general public. Premature deaths had also risen sharply, from 34.2 percent from NCDs take members of the labor in 2004 to 48.2 percent in 2012. Trends force away from their work at the height in physical measurements (for example, of their productivity, leaving families and body mass index) remained at a constantly children unsupported at a very early age. high level over time. 11 STEPS is abbreviation of STEPwise approach to surveillance. It is a simple, standardized method for collecting, analyzing and disseminating data in WHO member countries. There are currently two primary STEPS surveillance systems, the STEPwise approach to risk factor surveillance and the STEPwise approach to stroke surveillance. 12 Impaired fasting glycaemia (IFG) is a type of prediabetes, in which a person’s blood sugar levels during fasting are consistently above the normal range, but below the diagnostic cut-off for a formal diagnosis of diabetes. 28 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 1: Factors influencing health status for males and females, STEPS surveys 2004 and 2012 2004 2012 Risk factor Male Female Total Male Female Total Behavioral risk factors Current smoker (%) 46.2 14.3 29.8 46.4 13.4 29.3 Average daily consumption of <5 servings 91.4 92.9 92.2 72.4 73.7 73.1 of fruit and/or vegetables (%) Low levels of physical activity (1% with <600 Metabolic Equivalent of Task 33.3 53.7 43.9 15.1 31.7 23.7 minutes per week) Physical measurements Mean body mass index (BMI)(kg/m²) 31.7 34.9 33.3 31.3 34.8 33.1 Overweight (% with BMI ≥25) 89.2 94.9 92.1 87.3 94.0 90.7 Obese (% with BMI ≥30) 60.7 76.3 68.7 57.2 77.6 67.6 Overweight or obese (%) 89.2 94.9 92.1 87.3 94.0 90.7 Mean waist circumference (cm) 103.4 105.2 .. 103.3 106.7 .. Mean waist to hip ratio (cm) .. .. .. 0.9 0.9 0.9 Raised blood pressure (%) 26.5 19.9 23.1 28.2 27.1 27.6 Biochemical markers Impaired fasting glycaemia (%) 20.0 13.8 17.1 23.9 23.8 23.8 Raised fasting blood glucose (%) 16.3 16.6 16.4 29.7 38.6 34.4 Raised total cholesterol (%) 66.1 34.2 49.7 49.3 48.2 48.8 Source: 2004 STEPS survey (MOH and WHO, 2012); 2011/2012 STEPS survey (MOH and WPRO, 2013); The Kingdom of Tonga Health System Review, Health Systems in Transition, Vol. 5, No. 6, 2015. 29 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Tobacco consumption Tonga is one of the PICs with high also important to note that the prevalence tobacco consumption prevalence. As of tobacco use among school children shown in Table 2, around 29.3 percent of has increased. According to the latest the Tongan population smoke tobacco, WHO data, the prevalence of tobacco with 26.7 percent smoking on a daily use among school children ages 13–15 basis. Approximately half of Tongan is around 35 percent, with 44.9 percent adult males smoke. While there is no among male students and 28 percent significant change in the prevalence of among female students. The prevalence tobacco use between 2004 and 2012, the of tobacco use among female students average age of starting smoking fell in is significantly higher than that of adult 2012. Most smokers in Tonga in 2004 and females. If this continues, the overall 2012 smoked manufactured cigarettes, prevalence of tobacco use among Tongan and each smoker consumed approximately adults may rise in the future.14 11 sticks per day. Prevalence of tobacco consumption in Tonga compared with other The number of cases of lung and (selected) PICs is shown in Figure 4. tracheal cancer has rapidly increased in the last 30 years, largely due to the Continuously high tobacco high prevalence of smoking. The World consumption and the rise in smoking Tobacco Atlas estimates that 8 percent of among females suggest that female deaths and 7 percent of male deaths tobacco control and prevention in Tonga are due to tobacco,15 and the strategies used in the past may not number of admissions for chronic obstructive have been as effective as hoped.13 It is pulmonary disease (COPD) is rising. 13 The Government of Tonga has three different strands of tobacco control activity: its national taxation programme, bespoke national tobacco campaigns targeting issues such as second-hand smoking, and Ministry of Health-administered efforts such as quitline services for those willing to give up smoking. 14 WHO, Report on the Global Tobacco Epidemic, 2017. 15 The Kingdom of Tonga Health System Review, Health Systems in Transition, Vol. 5, No. 6, 2015. 30 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 2: Prevalence of tobacco use, 2004 and 2012 Tobacco use for adults ages 25–64 years Year Total Males Females 2012 29.3 46.4 13.4 Current smoker (%) 2004 29.8 46.2 14.3 2012 26.7 42.1 12.4 Current daily smoker (%) 2004 26.4 41.5 12.2 For those who smoke tobacco daily 2012 18.6 17.5 22.2 Average age started smoking (years) 2004 19.7 18.7 22.7 2012 85.2 81.8 95.9 Percentage of daily smokers smoking manufactured cigarettes 2004 82.5 82.8 81.4 2012 11.3 12 9.2 Mean number of manufactured cigarettes smoked per day (by smokers of manufactured cigarettes) 2004 12.9 13.3 11.6 Source: 2004 STEPS survey (MOH and WHO, 2012); 2011/2012 STEPS survey (MOH and WPRO, 2013); Hufanga, S., Comparative Effectiveness of Interventions to Control Non-Communicable Diseases in the Kingdom of Tonga, Draft Ph.D. Dissertation, University of New South Wales, 2018. Figure 4: Prevalence of tobacco consumption in the Pacific in 2016 Source: WHO Global report on trends in prevalence of tobacco smoking 2000–2025. 31 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Unhealthy diet Diet in Tonga, which plays an important a high consumption of unhealthy imported role in Tongan culture and in fulfilling and processed products. A recent familial and traditional responsibility, study examined household income and has changed over time. In the past, expenditure data in five countries in the Tongans consumed locally grown and Pacific – Kiribati, Samoa, Solomon Islands, produced products, and strict specifications Tonga, and Vanuatu. The study found existed as to the appropriate food types that Tongans spent 42 percent of their for different traditional events. However, expenditure on unhealthy food – due to globalization and expanding significantly higher than the other four international trade, imported and countries. Tongans were also the highest processed foods have become cheaper spenders on processed food, and spent and even more affordable than many over 50 percent of their expenditure on local products. Some of them have imported and non-traditional food (see become very popular, e.g. corned beef, Figure 5).16 salted beef, mutton flaps, turkey tails, sausages and sugar-sweetened beverages Unlike many other PICs, Tonga’s food (SSB). Hence, food consumption patterns expenditure was primarily on imported have changed, with Tongans consuming and mostly unhealthy cuts of high quantities of imported and processed meat, particularly mutton flaps and foods, which contain high amounts of processed meat. While other Pacific calories, fats, salt, and sugar. Small Islanders spent most on rice, Tongans businesses that sell affordable fast food spent most on mutton, mostly in the form of have also become popular in recent high-fat mutton flaps (Table 3). Spending on years, resulting in people consuming mutton is nearly twice as high as fewer home-cooked meals. poultry and cooked meat.17 Mutton flaps, considered high in animal fat, have been While more imported and processed identified in the 1992 Nutrition Survey as foods are consumed, Tongan society the most frequently consumed meat; 31 strongly maintains the tradition that percent of the study population consumed associates big meal sizes and plentiful mutton flaps every day, and 27 percent food with greeting each other and consumed it three or more times a week.18 embracing social relationships in family and the community. This has resulted in 16 Sahal S. et al, Trade as a structural driver of dietary risk factors for non-communicable diseases in the Pacific: an analysis of household income and expenditure survey data, 2014. 17 Sahal S. et al, Trade as a structural driver of dietary risk factors for noncommunicable diseases in the Pacific: an analysis of household income and expenditure survey data, 2014. 18 Hufanga, S., Comparative effectiveness of interventions to control non-communicable diseases in the Kingdom of Tonga, Draft Ph.D. Dissertation, University of New South Wales, 2018. 32 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 5: Expenditure on unhealthy food, imported food, processed food, Pacific Island Countries (selected), 2005―2010 Source: Sahal S. et al, Trade as a structural driver of dietary risk factors for noncommunicable diseases in the Pacific: an analysis of household income and expenditure survey data, 2014. While Tongans consume large Tongans in 2012 consumed only two amounts of fatty meat, consumption servings of vegetables per day. It is of fruits and vegetables is significantly important to note however that there was below WHO recommendations. As significant reduction among Tongans shown in Table 4, most Tongans do not who ate fewer than five servings consume fruits and vegetables on a daily of fruit and/or vegetables on average per basis. Fruits are consumed three days per day – from 92.2 percent in 2004 to only week, and vegetables are consumed four 73.1 percent in 2012. days per week. While WHO recommends three servings of vegetables per day, 33 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 3: Share of food expenditure in selected Pacific Island Countries, 2005–2010 Share of food Top 3 Share of food Top 3 Share of kcal Share of kcal expenditure products expenditure on products intake per item intake per item Country on item (% of food item out of total (kcal pcae/ out of total out of total kcal out of total expenditure) food expenditure day intake) imported foods intake imported foods Rice 43% 23% Rice 48% 34% Kiribati Sugar 25% 13% Sugar 27% 20% Flour 6% 3% Flour 11% 8% Rice 57% 17% Rice 53% 10% Solomon Noodle 13% 4% Flour 13% 3% Islands Sugar 7% 2% Sugar 12% 2% Rice 40% 12% Rice 34% 6% Vanuatu Tinned tuna 10% 3% Sugar 22% 4% Bread 7% 2% Bread 12% 2% Rice 27% 11% n/a n/a n/a Samoa Margarine 20% 8% n/a n/a n/a Bread and 10% 4% n/a n/a n/a Noodles Mutton 19% 10% n/a n/a n/a Tonga Poultry 12% 6% n/a n/a n/a Cooked 10% 5% n/a n/a n/a meat Source: Sahal S. et al, Trade as a structural driver of dietary risk factors for noncommunicable diseases in the Pacific: an analysis of household income and expenditure survey data, 2014. Table 4: Fruit and vegetable consumption in Tonga, 2004, 2012 Measurements (population ages 25–64 years) Year Total Males Females Mean number of days/week fruit consumed 2012 3.4 3.3 3.4 Mean number of servings of fruit consumed on average per day 2012 1.8 2 1.7 Mean number of days per week of vegetables consumed 2012 4.2 3.9 4.4 Mean number of servings of vegetables consumed 2012 2.1 2 2.2 on average per day 2012 73.1 72.4 73.7 Percentage who ate fewer than five servings of fruit and/or vegetables on average per day 2004 92.2 91.4 92.9 Source: 2004 STEPS survey (MOH and WHO, 2012); 2011/2012 STEPS survey (MOH and WPRO, 2013); Hufanga, S., Comparative Effectiveness of Interventions to Control Non-Communicable Diseases in the Kingdom of Tonga, Draft Ph.D. Dissertation, University of New South Wales, 2018. 34 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Physical Inactivity Physical activity is commonly defined least 150 minutes of moderate-intensity as a combination of structured activity physical activity per week or 75 minutes of and exercise, lifestyle activity, and vigorous-intensity physical activity, or the spontaneous physical activity which equivalent combination of moderate and generally aims to improve fitness, vigorous-intensity physical activity performance, and health, and can achieving at least 600 MET minutes per provide a means of social interaction. week.20 It helps to maintain energy balance, reduce the risk of obesity, and improve general Compared to the situation in well-being.19 Physical activity is measured 2004, there has been significant by Metabolic Equivalent of Task (MET). improvement in the level of physical MET is the objective measure of the ratio activity in Tonga. The proportion of adults of the rate at which a person expends with low levels of activity (defined as <600 energy, relative to the mass of that person, MET minutes per week) decreased from while performing some specific physical 43.9 percent in 2004 to 23.7 percent in activity, compared to a reference, set by 2012, and the proportion of adults with high convention at 3.5 ml of oxygen per kilogram levels of activity (defined as ≥3,000 MET per minute, which is roughly equivalent minutes per week) increased from 35.5 to the energy expended when sitting quietly percent in 2004 to 52.4 percent in 2012. While (this is equivalent to a caloric consumption improvement in physical activity is observed of 1 kcal/kg/hour and is defined as one in both male and female adults, it is of note MET). The WHO global recommendation that only 36.3 percent of female adults on physical activity for health is to do at engaged in high levels of activity, while 85.8 percent of them did not engage in vigorous activity. Table 5: Physical activity measurements, 2004 and 2012 Measurements (population ages 25–64 years) Year Total Males Females 2012 23.7 15.1 31.7 Percentage with low levels of activity (defined as <600 MET-minutes per week) 2004 43.9 33.3 53.7 2012 52.4 69.7 36.3 Percentage with high levels of activity (defined as ≥3,000 MET minutes per week) 2004 35.5 48.3 23.6 2012 64.8 42.3 85.8 Percentage not engaging in vigorous activity 2004 75.8 58.1 92.3 Source: 2004 STEPS survey (MOH and WHO, 2012); 2011/2012 STEPS survey (MOH and WPRO, 2013); Hufanga, S., Comparative Effectiveness of Interventions to Control Non-Communicable Diseases in the Kingdom of Tonga, Draft Ph.D. Dissertation, University of New South Wales, 2018. 19 Mumuiheata, S., Nutrient Intake in the Kingdom of Tonga. Nuku’alofa, 2007. 20 Ministry of Health, Kingdom of Tonga Risk Factors STEPS Report, 2014. 35 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Harmful alcohol use There is a strong link between (men 19.7 percent; women 3.9 percent) alcohol and NCDs, particularly cancer, than individuals ages 45–65 years (men: cardiovascular disease, liver disease, 9.3 percent; women 0.5 percent). pancreatitis, and diabetes. The harmful use of alcohol and more specifically Heavy episodic drinking or binge heavy episodic drinking is related to many drinking is still a major concern in cardiovascular outcomes such as strokes Tonga. According to the STEPS survey and hypertension.21 According to Tonga’s 2012, 46.9 percent of the population STEPS survey 2012, the overall prevalence (majority male) would drink more than six of adult alcohol consumption (ages 25–64) standard drinks in any one day. There is in the last 30 days was estimated at 9.3 also anecdotal evidence of increased percent with considerable gender difference crime and road traffic accidents, with – 16 percent for male and 2.8 percent for 2018 recording the highest number of female. Furthermore, current drinkers are deaths from road traffic accidents – the more likely to be individuals ages 25–44 majority of which were alcohol related.22 Figure 6: Percentage of the population with overweight and obesity (≥25 kg/m²) in Pacific Island Countries Source: WHO, STEPS surveys in respective countries except Wallis (SPC-led survey similar to STEPS) and CNMI (hybrid survey led by PIHOA). 21 Parry, C. D., Patra, J., & Rehm, J. Alcohol consumption and non-communicable diseases: epidemiology and policy implications. Addiction (Abingdon, England), 106(10), 1718-24. 2011. 22 Parry, C. D., Patra, J., & Rehm, J. Alcohol consumption and non-communicable diseases: epidemiology and policy implications. Addiction (Abingdon, England), 106(10), 1718-24. 2011. 36 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Obesity Obesity is a medical condition in which Tonga has the second highest excess body fat has accumulated to an prevalence of overweight and obesity extent that it may impair health. People (BMI ≥25 kg/m²) of all PICs.24 Prevalence are generally considered obese when their of overweight (BMI ≥25 kg/m²) is over 90 body mass index (BMI), a measurement percent, while the prevalence of obesity obtained by dividing a person’s weight by (BMI ≥30 kg/m²) is 67.6 percent. According the square of the person’s height, is over to the STEPS survey 2012, 87.3 percent 30 kg/m2; the range 25–30 kg/m2 is defined of men and 94 percent of women were as overweight. The fundamental cause classified as either overweight or obese. of obesity and overweight is an energy As shown in Table 6, the prevalence of imbalance between calories consumed obesity has risen significantly among and calories expended – most commonly women – from an average weight of due to a combination of excessive food 73.9 kg in 1973 to 95 kg in 2004. The intake and lack of physical activity. Obesity corresponding increase for males is from and overweight increase the likelihood 79.1 kg to 95.7 kg over the same period. By of cardiovascular diseases (mainly heart 2012, 77.6 percent of women were obese disease and stroke), type 2 diabetes, while 57.20 percent of men were obese musculoskeletal disorder, and several – a difference of nearly 20 percentage types of cancer. Childhood obesity is points. Both obesity and overweight are associated with a higher chance of obesity, common at younger ages, with about 20 premature death and disabilities, percent of Tongan boys and girls being musculoskeletal disorder, and several obese.25 As there is no significant change types of cancer.23 in the prevalence of overweight and obesity between STEPS surveys 2004 and 2012, more effective interventions to tackle overweight and obesity are needed. Table 6: Body mass index measurements in Tonga, 2012 Measurements (population ages 25-64 years) Year Total Males Females 2012 33.1 31.3 34.8 Mean BMI 2004 33.3 31.7 34.9 2012 90.7 87.3 94.0 Percentage with BMI ≥25 kg/m² 2004 92.1 89.2 94.9 2012 67.6 57.2 77.6 Percentage with obesity (BMI ≥30 kg/m²) 2004 68.7 60.7 76.3 Source: 2004 STEPS survey (MOH and WHO, 2012); 2011/2012 STEPS survey (MOH and WPRO, 2013); Hufanga, S., Comparative Effectiveness of Interventions to Control Non-Communicable Diseases in the Kingdom of Tonga, Draft Ph.D. Dissertation, University of New South Wales, 2018. 23 Obesity and Overweight Key Facts, World Health Organization. 2018. 24 American Samoa is the PIC with highest prevalence of overweight and obesity at 93.5 percent. 25 The Kingdom of Tonga Health System Review, Health Systems in Transition, Vol. 5, No. 6, 2015. 37 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The government’s NCD response In the past decade Tonga has joined • Improving the efficiency and impact of other PICs in recognizing the NCD the health sector for prevention and early crisis, its impacts on economic treatment. development, and in mobilizing a response to the crisis. In response to • Strengthening the evidence case for the UN-convened High-level Meeting better investment planning and program on the Prevention and Control of NCDs effectiveness. in May 2011, the Pacific Island Forum Leaders meeting declared in September Tonga became the first PIC to launch 2011 that the Pacific is currently a National Strategy for the Prevention experiencing an NCD crisis. It was stated and Control of NCDs and set that “75 percent of all adult deaths in the up structures and processes for Pacific are due to NCDs, the majority of multisectoral collaboration. The National whom are in the economically active age NCD Committee (NNCDC) was also set bracket and that many more times this up, chaired by the CEO of the Ministry of number suffer severe side effects that Health and comprising ministers from the undermine their capacity to contribute Ministry of Education, Ministry of Internal further to economic development”.26 At Affairs, Ministry of Agriculture, Food, Forests, the direction of the Forum of Economic and Fisheries27, and the Ministry of Finance Ministers, the NCD Roadmap was prepared and National Planning. It also includes the to guide PICs in the most effective use of Police Commissioner, CEO and Chairman resources to prevent and control NCDs. of the Tonga Health Promotion Foundation, In 2014, at the inaugural Joint Forum and representatives from the National Economic and Pacific Health Ministers Forum of Church Leaders, from the business Meeting, the ministers committed to five sector, and from civil society. The NNCDC key measures: has been supported by four advisory committees for Healthy Eating, Physical • Strengthening tobacco control, including Activity, Alcohol, and Tobacco. The National raising the excise duty to 70 percent of the Strategic Plan for NCDs is aligned with the retail price of cigarettes. Tonga Strategic Development Framework (TSDF II 2015–2025) which calls for more • Considering an increase in taxation of collaborative efforts across government alcohol products as a way of reducing ministries, civil society organizations and harmful alcohol consumption. faith-based organizations. • Reducing consumption of food and drinks directly linked to obesity, heart disease and diabetes such as sugar- sweetened beverages (SSBs), salty and fatty food (including through tax measures). 26 Tonga, National Strategy for Prevention and Control on Non-Communicable Diseases 2015-2020, 2016. 27 This ministry was later separated into Ministry of Agriculture, Forestry, and Food; and Ministry of Fisheries 38 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The Government of Tonga has gradually products, including mutton flaps), sugar- introduced taxation measures to sweetened food (e.g. ice cream, chocolates), influence tobacco consumption in instant noodles, and SSB, with the aim recent years. For example, in mid-2013, of increasing the price of these products, excise tax on imported cigarettes was and in turn reducing consumption. increased from T$210 per 1,000 cigarette sticks to T$250 – and T$200 to T$238 for The almost 50 percent increase in locally manufactured cigarettes. At the excise tax on imported cigarettes in same time the tobacco concession for July 2016 was particularly significant. inbound travelers was reduced from 500 Unlike the small, incremental increases cigarettes to 250 cigarettes. From July of previous years, the July 2016 hike 2016 Tonga further increased taxes on resulted in the excise tax on imported cigarettes and alcohol, and imposed new cigarettes rising from T$255 to T$380 per excise taxes on a number of unhealthy 1,000 cigarette sticks (full details on the food and beverage products, including tax increase and its effects on the prices selected fatty foods (in particular meat of different products are in the Findings section). A snapshot of tobacco control interventions appears as Box 1. Box 1: Government of Tonga tobacco control activities: a snapshot The Government of Tonga undertakes three main tobacco control activities, and across these activities there is a varying degree of success. Taxation policy Tonga’s national taxation policy, which started in 2013, is among the strongest tobacco tax policies of any Pacific country, and applies excise tax, consumption tax and import duty on many tobacco products. As in other countries, Tonga’s cigarette excise tax is the most important for determining the tax incidence and price of tobacco products and is imposed at the import and production level in the form of a specific tax. Consumption tax is imposed at import, production, wholesale and retail levels. Import duty has been applied on imported tobacco products since FY2015–6. The most significant intervention came in July 2016 as the government increased the excise tax on imported cigarettes by nearly 50%. However, the positive impacts of Tonga’s tobacco tax policy have been diluted by the availability of cheaper substitutes. There has been a shift from more expensive brands of cigarettes to cheaper brands, particularly locally manufactured cigarettes that are subject to less tobacco tax. But more importantly, there was a big shift from manufactured cigarettes to the untaxed Tapaka Tonga. Today, most tobacco users in Tonga smoke Tapaka Tonga. National tobacco control campaigns Tonga’s Ministry of Health and the Tonga Health Foundation have to date run a single, two-phase campaign. The first phase was a mass-media campaign in June 2016 entitled ‘Cigarettes are Eating Your Baby Alive’, and highlighted the harms of second-hand smoking. It targeted the population ages 18–64 years. A campaign evaluation revealed that there was a significant increase in the proportion of people avoiding places where they would be exposed to second-hand smoke, 39 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga in addition to greater awareness of the consequences of second-hand smoking. Nevertheless, smokers were shown to perceive that their right to smoking was a ‘human right’, of equal importance to the right to public health. The second phase of the campaign in August 2017 aimed to increase awareness and knowledge of the consequences of cigarette smoking using the strapline ‘Eating You Alive’, and was linked to the Ministry of Health’s Quitline services. The evaluation of this phase of the campaign found that those reached by the campaign felt it: • Taught me something new (89.7%) • Is relevant to me (88.7%) • Encouraged me to discuss quitting with friends and family (87.6%) • Motivates smokers to quit (85.9%) • Increase the likelihood of smokers calling the Quitline (83.1%) The campaign also identified two key factors motivating smokers to quit – factors that increased in importance after the campaign: • Being healthy for family, from 30.8% (pre-campaign) to 36.6% (post-campaign) • My own health or fitness, from 20.7% (pre-campaign) to 30.1% (post-campaign) The evaluation also suggested that any future anti-tobacco campaign must include promoting a call to action by incorporating Quitline service information in all campaign advertisements. Enforcement and compliance Efforts to enforce and encourage compliance with tobacco-control measures such as preventing smoking in public and prohibited places, as well as providing quitline services for those who are willing to give up smoking, are mostly administered and delivered by the Ministry of Health. Enforcement measures include visits by the Tobacco Control Unit Enforcement Team in order to conduct compliance checks of tobacco outlets and smoke free areas (SFA). The team visits locations such as shops, schools, town and church halls, kava clubs and bars. One such visit to the community of Vava’u involved the team visiting 54 shops. They found that 16 shops were selling to minors and that none of the shops had signage explaining that sales to minors were illegal; 12 shops were illegally displaying tobacco; and 10 shops lost or misplaced their ‘smoking kills’ sign. Most of the halls and kava clubs still allowed people to smoke inside. Clarification on the new Act was provided to owners and members, highlighting the infringement notice and penalty components of the Act. Verbal warnings were given as the outlets had not been fully aware of the Act, and in turn, proprietors gave assurances that they would comply with the Act in the future. 40 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Enforcement activity needs to be strengthened and conducted, and training on the new Act is needed to raise awareness of how to comply on behalf of both employees and owners of shops, kava clubs, bars, and halls. Awareness raising is also needed among the general public. Summary report on quitline services provided during each anti-tobacco campaign as from 2016 –2018 Year Campaign Number of Number Number of Number Number of phases Callers of people people who of people people lost due to registered quit still on the change of phone with Quitline program numbers/migrate 2016 Phase 1 167 171 69 123 48 2017 Phase 2 153 138 46 121 17 2018 Phase 3 207 167 51 143 24 2019 Phase 4* Total 527 476 166 387 89 *Phase 4 will be launched in May 31, 2019 as highlight of the World No Tobacco Day celebration in Tonga Tonga has implemented a number of development partners, the Ministry of multi-sectoral interventions to address Fisheries also has focused on the special NCDs, and its tax interventions in management area program (SMA) which particular have been recognized by aims to improve the health of reefs, marine WHO. As well as becoming one of the first and sea organisms, and ultimately result in countries in the Pacific to introduce taxes the provision of healthier sources of food. on tobacco, alcohol, and unhealthy food and beverages, Tonga also abolished In addition to its taxation measures, tax on imported fruit and vegetables in the Government of Tonga has also 2016, with the aim of increasing their strengthened NCD-focused primary consumption. The government has also health care services. Through the Ministry strengthened the implementation of the of Health it has introduced NCD nurses WHO Framework Convention on Tobacco to provide complementary NCD services Control, launching mass media tobacco in rural health services, including health campaigns and cessation services in promotion activities, screening and follow-up, recent years, and joining forces with and care of NCD cases. The government the church and communities to expand has achieved WHO recognition for the community-based physical activity work done in relation to tobacco taxation, initiatives. The Ministry of Fisheries has and for interventions to increase women’s introduced a new policy that requires and girls’ physical activity through World foreign fishing vessels to offload a certain No Tobacco Day awards in 2014 and the amount of fish for sale through the Kau Mai Tonga netball project which won a government, to increase community healthy island recognition award in 2013. access to fish as part of their diet. With assistance from Tonga Health among other 41 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The Government of Tonga is committed In most countries sin taxes on to strengthening monitoring and traditional items such as tobacco, evaluation activities for its NCD alcohol, and sugary beverages help response. This study forms part of reduce consumption. However, attempts the government’s efforts to understand to tax unhealthy food products have not the progress, impacts, challenges, been as comprehensive as tobacco and lessons learned from its tax policy and alcohol taxes, as their design and implementation. However, as these application are usually more complex, policies have been implemented alongside given the many potential substitutes other health promotion activities and available. campaigns described above, it is not possible to accurately attribute the impacts Tonga’s unique dietary situation (where to solely one policy or activity. imported products account for a large part of the foodstuffs available) and Making the most of ‘sin taxes’ pressing health conditions have pushed the government to be ambitious Taxes have long been used to serve in taxing unhealthy food products, and society, encouraging behaviors that the it now has one of the most extensive government favors with tax incentives lists of taxed food items in the world. and credits, while discouraging For this reason the monitoring and behaviors it considers undesirable evaluation of Tonga’s excise tax policy on with increased or additional taxes. ‘Sin unhealthy food products is of huge value to taxes’ are designed to discourage harmful the scientific community, as it will generate behaviors and generate revenues. important evidence on the effectiveness of Traditionally, sin taxes were designed taxes from which other countries can learn. to control or curb vices perceived as undesirable, such as tobacco and alcohol use, and gambling. As our economies evolve, so too have these perceived vices, and governments (states and localities) have begun to specifically tax ‘sin’ items such as tobacco, sugary beverages, plastic bags and other potentially harmful products. 42 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Methodology 43 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga This study employed a variety of consent was acquired before the interviews, assessment tools and data sources and interviewees could withdraw at any to obtain feedback from various stage. The endline survey was conducted stakeholders involved in Tonga’s tax in November and December 2017. With the policy implementation. These included assistance of the Department of Statistics, (i) government administrative data, the sample surveyed was distributed to including import volumes, revenues, and ensure it was representative of households the consumer price index; (ii) baseline on the main islands of Tongatapu and and endline household surveys, focusing Vava’u, which make up over 85 percent of on the impact of NCD tax on consumption Tonga’s population. behaviors; (iii) retail surveys focusing on the impact of NCD tax on prices; (iv) tobacco Sampling design taxation simulation modeling (TaXSiM); (v) a qualitative study, which includes An initial sample of 1,400 individuals focus group discussions and in-depth (respondents) ages 18–65 was chosen interviews. As required by the Ministry of as household representatives to Health, the World Bank team submitted undertake the Tonga NCD Taxation the research proposal with a methodology Household Baseline Survey 2017. Table and implementation plan to Tonga’s 7 provides information of total sample size National Health Ethics and Research and island-wise distribution.28 Committee (NHERC) for review. The research proposal was approved by Table 7: Total sample size of baseline data NHERC in mid-2017, allowing the Number of Required Island division quantitative surveys to take place in households sample 2017–2018 and the qualitative study to be Tongatapu 12,953 1,000 carried out in 2018. Vava'u 2,715 400 Household surveys Total 15,668 1,400 Baseline and endline household Source: Baseline survey data (2017) surveys were conducted to assess the pre-tax/post-tax consumption The sampling in each area was behaviors and to have preliminary then undertaken using a two-stage understanding of potential substitutes. randomized sampling process. The first The baseline household survey was stage involved the random selection of conducted in May 2017 to understand the required sample from each village the consumption behaviors of consumers using Probability Proportional to Size (PPS) before the July 2016 taxation policy sampling, with number of households intervention (asking consumers to recall) as the size. For the second stage, a list of all and before the July 2017 taxation policy households in every village included in the intervention (present behaviors). The frame was taken from the Tonga National questionnaires were carried out in the Population and Housing Census, 2016. Tongan language. Consent forms/verbal ‘Head of households’ lists for all villages 28 The sample was selected independently within each of the two target areas. Firstly, extremely remote areas were removed from the frame (and thus not offered for selection) as it was considered too costly to cover these areas. These areas only represented about 2.5 percent of Tonga’s total population, so the impact of their removal was minimal. 44 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga were sorted alphabetically and then a systematic sampling was used to draw out each required sample from each village with household names, including the names of household members who were currently smoking. This gave every household equal chance of getting selected. GPS coordinates and maps were provided to help enumerators locate the households. Enumerators interviewed the head of household or his/her representative. If that person did not smoke, then the name of the person who smokes given on the list would be called to answer the tobacco-related questions. The same process was applied to questions related to alcohol consumption. The above methodology was also used in the collection of endline data. The sample for the endline survey 2017 was drawn from the Household List of the most recent Population and Housing Census, 2016. The sample was designed to cover 9 percent of all households in Tonga excluding the Ongo Niuas island and some small islands due high transport cost. The total sample included 1,600 households – Table 8 shows the distribution between islands.29 Table 8: Total sample size of endline data Number of Required Island/ division households sample Tongatapu 12,944 1,000 Vava’u 2,503 400 Ha'apai 1,179 114 'Eua 885 86 Total 17,511 1,600 Source: Baseline survey data (2017) 29 As the Department of Statistics is interested in understanding the consumption behaviors in the other two island groups, Ha’apai and ‘Eua were added to the survey, as there are possibilities that the government may conduct follow-up surveys in these islands in the future. 45 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Household characteristics of Household characteristics of baseline data endline data The original dataset contains 1,402 The original dataset contained 1,604 household observations. After dropping household observations. After omitting observations with missing values, there observations (or data points) with missing are 1,395 household observations to values, 1,573 households were included be used for subsequent analysis. There for data analysis. It was possible that are 996 household observations from more than one member in a household Tongatapu; the remaining 399 household was interviewed if various household observations are from Vava’u. Around 65 members smoked or used alcohol. Thus, percent of the respondents are male, and the actual number of persons interviewed the average age of respondents is 39.5. was more than number of household Their average monthly earnings total observations. There are 1,475 individual T$996.60. The average monthly earning observations from Tongatapu; there are for a male is T$1,005, slightly more than 609 individual observations from Vava’u; Tongan females who earn approximately 131 individual observations are from T$980 per month. Household sizes of both Ha’apai; and the remaining 86 individual sexes are slightly above six members observations are from ‘Eua. Table 9 per household. However, there is a provides descriptive statistics on great variation in number of household respondents’ socio-economic background. members, ranging from 1 to 24. When Around 59 percent of respondents are examining marital status, 70.18 percent of male, and the average age of respondents respondents are married and 23.58 percent is 43 years. Their average monthly earning are never married. The proportion of never- is T$1,285. The average monthly income married males is 6 percent higher than for men is T$1,321, which is slightly higher females. Island-wise descriptive statistics than the monthly income from women are shown in Table 20, Annex 1, and (approximately T$1,233 per month). Table 9 provides descriptive statistics on Average household size is slightly above respondents’ socio-economic background. six members per household. Island-wise descriptive statistics are shown in Table 22, Annex 2, and Table 9 provides descriptive statistics on respondents’ socio-economic background. Approximately 80 percent of households in the baseline survey participated in the endline survey.30 30 The data were not originally designed to be longitudinal given the logistics limitation. However, the team matched the households in the baseline and endline surveys and found that approximately 80 percent of the households in the baseline survey participated in the endline survey. This is considered pseudo-panel data. The team matched the households from both surveys using a three-step matching process: Step 1: matching observations by village names; Step 2: matched again by gender in the same villages; Step 3: for the same gender of the same village, we matched the data by the corresponding Wealth Index at two digit level. 46 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 9: Descriptive statistics of baseline and endline data Baseline Endline Variable Male Female Overall Male Female Overall Gender 65.73% 34.27% 100.00% 58.89% 41.11% 100.00% Age 39.73 38.85 39.43 43.11 42.12 42.70 Earning 1,005.32 979.88 996.60 1,321.12 1,233.04 1,285.11 Number of household members 6.14 6.69 6.33 6.0 6.4 6.19 Education No formal schooling 0.11% 0.00% 0.07% 1.47% 0.27% 0.98% Primary school 2.18% 2.30% 2.22% 0.74% 0.53% 0.65% Lower secondary 26.28% 15.69% 22.65% 23.66% 24.20% 23.88% Upper secondary 54.53% 65.90% 58.42% 59.02% 61.97% 60.23% Technical and vocational 11.12% 9.00% 10.39% 9.67% 7.71% 8.87% Tertiary 5.23% 6.49% 5.66% 5.43% 5.32% 5.39% Other 0.55% 0.63% 0.57% - - - Marital status Never married 25.63% 19.67% 23.58% 13.81% 12.50% 13.28% Currently married 69.57% 71.34% 70.18% 82.32% 76.33% 79.87% De facto or consensual marriage 0.44% 0.42% 0.43% 0.18% 0.67% 0.38% Separated 0.98% 1.46% 1.15% 0.74% 1.46% 1.03% Divorced 0.44% 1.88% 0.93% 0.74% 1.99% 1.25% Widowed 2.94% 5.02% 3.66% 2.21% 7.05% 4.19% Other 0.00% 0.21% 0.07% - - - Occupation Employer 3.16% 0.42% 2.22% 2.30% 1.20% 1.85% Self-employed 13.96% 15.69% 14.55% 9.94% 13.56% 11.43% Employee, working for wages (type 1) 9.60% 5.02% 8.03% 14.27% 6.38% 11.04% Employee, working for wages (type 2) 24.65% 9.62% 19.50% 28.08% 10.20% 20.78% Producing goods for own and/ or family 23.23% 3.77% 16.56% 25.69% 12.63% 20.30% Unpaid family worker (family business) 5.56% 0.63% 3.87% 6.80% 6.50% 6.69% Unpaid family worker 1.74% 12.76% 5.52% 0.55% 2.66% 1.41% (help with basic household activities) Volunteer work 0.98% 0.00% 0.65% 1.01% 0.13% 0.65% Student ― full time 2.18% 1.88% 2.08% 0.46% 0.27% 0.38% Student ― part time 0.44% 0.63% 0.50% 0.74% 0.13% 0.49% Home duties 12.43% 49.16% 25.02% 9.58% 46.01% 24.48% Retired 1.53% 0.21% 1.08% 0.46% 0.27% 0.38% None ― did not pursue any activity 0.44% 0.21% 0.36% - - - Physically/mentally disabled 0.11% 0.00% 0.07% 0.09% 0.00% 0.05% Total number of observations 917 478 1,395 1,355 946 2,301 Source: Baseline and endline survey data (2017). 47 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga It is important to note that analysis is The next step is to calculate the mainly carried out at household level. socio-economic index (SI) for each Hence, data representativeness should household. This is done by the formula be examined at the household level. To shown below. In this formula, hsi is assess the representativeness of the data, the socio-economic score for each of the team assessed two issues. The first the households obtained from the PCA relates to distribution of households among exercise above. Min(hsi) and Max the four islands. The second relates to (hsi) represent minimum and maximum gender distributions. By and large, technical socio-economic scores respectively. The examination has revealed that the value of SI will be between 0 to 1; a higher household endline data structure resembles SI score denotes higher socio-economic that of the Household Income and status of the household. These values Expenditure Survey 2016 of Tonga. This are then divided into quintiles for further provides the team with a certain level of data analysis. Further details of the method comfort in using the data for analysis. can be found in Annex 4. Detailed analysis of this issue appears as Annex 3. Data analysis Data were analyzed using descriptive Retail surveys statistics such as means, median, standard deviation, minimum and The purpose of the retail surveys is maximum value. Analyses are done at to understand the prices of selected both aggregate level, island-wise and by products affected by the increase in socio-economic background. To provide excise tax/import duty, as well as the analysis beyond the mean of the data, products affected by the exemption of differences in socio-economic background consumption tax. Baseline surveys were have also been analyzed by applying the conducted in May 2017 to understand method from Ebert and Welsh (2004) and the price of products before the July OECD (2008). It begins with selecting 2016 taxation policy intervention (asking a set of socio-economic variables that the retailers to recall) and before the would capture key characteristics of 2017 taxation policy intervention (actual a household. We use types of wall price at that time). The implementation materials, source of lighting, types of fuel of the endline survey was delayed by used for cooking, types of toilet facilities, Cyclone Gita. It was implemented in car ownership, presence of a refrigerator, Vava’u in March 2018, and in Tongatapu washing machine, desktop or laptop in June 2018. The surveys were conducted computer, and earnings and education of in 35 shops in Tongatapu and 12 shops head of household as a set of information in Vava’u. The information from the to capture households’ socio-economic retail surveys helped fill in the data status. These variables are then analyzed gaps in the consumer price index, which by Principle Component Analysis (PCA) covered only limited products and brands. to obtain a statistical model. The model will contain information of relative importance Qualitative study of these variables in representing a household’s socio-economic status. A qualitative approach was used to This model is used to predict a socio- ascertain more in-depth perspectives economic score for each household. and opinions. A qualitative study tool 48 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga including focus group discussions and were organized to elicit feedback from key informant interviews was developed households and communities with by the study team. It was then peer different characteristics, taking into reviewed by local and international account gender, urban or rural, as experts to support the assessment, as well as well as socio-economic status. Given to triangulate and provide complementary the sensitivity in categorizing community evidence for the quantitative study. members into poor and wealthy groups, the research team used well-off and less Semi-structured questionnaires were well-off categories based on the following developed based on findings from the four criteria: (i) sustainable weekly/monthly baseline household survey, the results income from paid job (government/ of which were presented to stakeholders non-government); (ii) sustainable income in a workshop in September 2017. The from farming and/or handicrafts; (iii) study tools included key informant interviews sustainable income from fishing; (iv) with policymakers and focus group sustainable income (remittance) from discussions with communities. The detailed families and friends overseas. Participants qualitative tools and methodology were were categorized as well-off if they matched approved by Tonga NHERC. at least two of the above four criteria, and were considered as participants Focus group discussions with sufficient financial resources to meet basic needs in their community. Table 10 Focus group discussions (FGDs) lists the FGD categories. Table 10: Focus group discussion categories Pre-test focus group No. Place Category Socio-economic status Sex No. Attended 1 Pre-test (food consumption) Food Not well-off Male 4 2 Pre-test (tobacco) Tobacco Not well-off Female 4 Total 8 Focus group discussion No. Vava’u Category Socio-economic status Sex No. Attended 1 Neiafu (Urban) Food Well-off Male 5 2 Neiafu (Urban) Food Not well-off Female 6 3 Neiafu (Urban) Food Not well-off Male 8 4 Neiafu (Urban) Food Well-off Female 9 5 Neiafu (Urban) Tobacco/Alcohol Combined due to limited numbers Female 7 6 Neiafu (Urban) Tobacco/Alcohol Combined due to limited numbers Male 5 7 Tu'anekivale (Rural) Food Well-off Female 9 8 Tu'anekivale (Rural) Food Not well-off Female 2 9 Tu'anekivale (Rural) Tobacco Combined due to limited numbers Female 5 10 Tu'anekivale (Rural) Tobacco Not well-off Male 7 11 Tu'anekivale (Rural) Tobacco Well-off Male 6 12 Ofu (Rural – outer island) Food/Tobacco/Alcohol Combined due to limited numbers Female 9 13 Ofu (Rural – outer island) Food/Tobacco/Alcohol Combined due to limited numbers Male 13 Total 91 49 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 10: Focus Group Discussions Focus group discussion categories (continued) No. Tongatapu Category Socio-economic status Sex No. Attended 1 Houmakelikao (Urban) Alcohol Well-off Male 6 2 Houmakelikao (Urban) Alcohol Not well-off Female 4 3 Ma'ufanga (Urban) Alcohol Well-off Female 4 4 Houmakelikao (Urban) Food Not well-off Combined (Youth) 16 5 Fasi (Urban) Alcohol Not well-off Male 7 6 Ma'ufanga (Urban) Food Well-off Male 11 7 Ma'ufanga (Urban) Food Well-off Female 6 8 Houmakelikao (Urban) Food Not well-off Female (2 males) 10 9 Ha'asini (Rural) Tobacco Well-off Female 6 10 Ha'asini (Rural) Food Well-off Male 5 11 Ha'asini (Rural) Food Well-off Female 5 12 Ha'asini (Rural) Food Not well-off Male 8 13 Ha'asini (Rural) Food Not well-off Female (2 males) 10 14 Ha'asini (Rural) Tobacco Well-off Male 6 15 Ma'ufanga (Urban) Tobacco Not well-off Male 7 16 Ma'ufanga (Urban) Tobacco Well-off Male 6 17 Ma'ufanga (Urban) Tobacco Not well-off Female 5 18 Ma'ufanga (Urban) Tobacco Well-off Female 5 19 Ha'asini (Rural) Alcohol All Male 12 20 Ha'asini (Rural) Alcohol All Female 10 Total 149 Training of fieldwork staff and pre-test February 2018, for a period of three weeks of the qualitative materials at Tongatapu followed immediately by two weeks at Vava’u. Fieldwork design A team of four research assistants was was formulated in consultation with the recruited to support implementation Government Statistics Department as well in the field, working with two Tonga- as the Ministry of Internal Affairs. It was based senior consultants under overall decided that Ma’ufanga village would guidance and supervision of the World represent the urban residences and Bank’s Task Team Leader. Fieldwork Ha’ssini would represent the rural village staff received general qualitative training of Tongatapu. It was also decided that and specific sessions on the qualitative Neiafu would represent urban households study questionnaire. At the end of their in Vava’u, Tu’anekivale rural areas, and training, a pre-test session was carried then the Island of Ofu would represent out in two groups (tobacco and food the smaller inhabited islands of Vava’u. consumption) on February 27, 2018. However, a Category 4 tropical cyclone Implementation plan and variation hit Tongatapu on February 12, 2018, causing severe damage and affecting The original plan was to conduct the the implementation of the qualitative fieldwork straight after the pre-testing study. The damage was so great that the of questionnaires in the last week of implementation plan had to be altered. This avoided bias in the focus group 50 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga requirements to support the qualitative study in their respective communities) was provided by the research team for Ministry of Internal Affairs’ administrative staff, and district and town officers. A total of 248 participants attended focus group discussions with 13 groups in Vava’u and 22 focus group discussions in Tongatapu, including two pre-test groups. Of the participants, 46 percent were male, 48 percent were female, and 6 percent were youth under the age of 25 years. Audio recordings of the focus group discussions were then translated and transcribed, and shared with the research team for analysis. Key Informant Interviews As part of the qualitative study, key informant interviews were conducted with policymakers in relevant government agencies as well as NGO stakeholders, church leaders, discussion as a result of this tragic event, community representatives, and as well as observing the government’s private sector representatives. The aim and the World Bank’s recommendations was to gain an in-depth understanding of to allow key government officials and the roles of different agencies and leaders communities to focus on recovery efforts in the design and implementation of the following the cyclone. The implementation NCD tax policy. Interviews also aimed started in March in two villages and to solicit their feedback on the impacts, on the island of Vava’u, which were not challenges, and lessons learned from directly affected. Following government the implementation of NCD tax policy. A advice, implementation of the qualitative semi-structured questionnaire was designed study in Tongatapu started in late to support the key informant interviews June 2018 and concluded in August and was approved by NHERC. The 2018 for both focus group discussions implementation of the key informant and key informant interviews. interviews was also affected by Cyclone Gita. Key informant interviews were planned Participants to be conducted during the same period as the focus group discussions. However, The research team worked closely with following advice of the government, staff of the Local Government Division implementation of the key informant of the Ministry of Internal Affairs to interviews was postponed to July and identify and recruit participants for the August 2018 in order not to disrupt the focus group discussions. An introductory post-Cyclone Gita recovery efforts. The session on the background of the overall list of key informants interviewed is in impact study (and in particular the logistics Annex 5. 51 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Tobacco taxation simulation Key limitations modeling (TaXSiM) Recall bias The Tobacco Tax Simulation (TaXSiM) model is an innovative tool that can be The World Bank team acknowledged the used to describe the current market limitation resulting from the recall bias and tax situation for cigarettes in a that affected the baseline household particular country or tax jurisdiction, survey. The team received the request and to forecast the impact of tax changes from the Government of Tonga at the end on final consumer prices, cigarette of 2016 to assess the impacts of the NCD consumption and government tax tax policy introduced in July 2016. The revenues. A particular strength of the model recall bias affected all the products subject is that it examines outcomes on a brand-wise to the excise tax increase in July 2016. basis, which highlights how different tax However, it does not affect assessment policies can affect different segments of of alcohol consumption behaviors, as the the tobacco market. The TaXSiM Tonga excise tax on alcohol was increased in model selects main cigarette brands July 2017, and not July 2016. which accounts for about 90 percent of Tonga’s cigarette market. The selected The team undertook several measures brands include main imported cigarette to minimize recall bias. First, the team brands (Pall Mall, Winfield, Long Beach) conducted an extensive literature survey and locally produced cigarette brands related to addictive consumption. It was (Palataisi, Halo). The TaXSiM Tonga model found that recall bias could significantly uses the market share of each brand to affect accuracy of consumers who consume calculate the weighted average price, the products sporadically. For those with and to weight average tax incidence, etc. high consumption intensity, however, the bias is small (McLaughlin et al., 1987; Data populated into TaXSiM included Krall et al., 1989; Giovannuci et al., 1993; volumes of locally produced and Gmel and Daeppen, 2007). Based on imported cigarettes, tax revenue the value reported in Figure 18, around including import duty, excise tax, 80 percent to 90 percent of smokers consumption tax, CIF and retail price – smoked daily, so any recall bias should all sourced from internal data provided not significantly influence results. by the Ministry of Customs and Revenue. Retail prices for FY2016–7 and FY2017–8 For alcohol and other consumer were sourced from retail surveys products, we followed the approach conducted under this study. Retail prices adopted by Patrick et al. (1994), of previous years were sourced from Fendrich et al. (2005), Weitkunat et local tobacco agents and consumer al. (2013) which uses a range-based price index records provided by Tonga’s response for interviewees rather Department of Statistics. It should be noted than asking them to provide an exact that the analysis results may not be robust quantity. This helped to increase the for the years for which officially recorded accuracy of the results. retail prices are not available. 52 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga When possible, we compared and poor households from household surveys triangulated household survey data with for a qualitative study is a sensitive issue Household Income and Expenditure in Tonga. The team later found a solution Survey data, imported data by working with staff of the Local Government (administrative data), retail survey Division of the Ministry of Internal Affairs, data, and results from focus group particularly district and town officers, in discussions. The results from this identifying and recruiting well-off and not revealed a high degree of consistency so well-off participants for focus group across different datasets, providing a discussions. The team explicitly avoided satisfactory level of comfort in the survey the use of the terms “poor” and “near-poor”. results. Participants were categorized as well-off if they complied with at least two of the Measuring FGD members’ four criteria, namely sustainable weekly/ socio-economic status monthly income from a paid job (government or non-government), sustainable income The research team applied a unique from farming and/or handicrafts, sustainable means of measuring FGD members’ income from fishing, and sustainable socio-economic status in Tonga – a income (remittances) from families and method the team strongly believes is friends residing overseas. The participants rigorous and takes into account the in non well-off groups were mainly constraints and observing country- unemployed. specific sensitivities. The team consulted the Department of Statistics in identifying poor/near-poor participants (the bottom 40 percent) and non-poor participants (the remaining 60 percent) for the qualitative study for the baseline and endline household surveys. However, the team was informed that identifying poor/near 53 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Findings 54 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Tobacco Price elasticity Tonga’s tobacco tax structure There is lower sensitivity of cigarette demand to the variation of cigarette Tonga has some of the strongest tobacco prices, with a 10 percent increase in tax policies of any Pacific country, and cigarette prices leading to a 1.1 percent applies excise tax, consumption tax and decrease in cigarette consumption. To import duty on cigarette products. As in estimate price elasticity of demand for other countries, Tonga’s cigarette excise cigarettes in Tonga 2011–17 we derived tax is the most important part of the tax the price elasticity of demand by the incidence and price of cigarette products following equation: and is imposed at the import and production level in the form of a specific tax. Consumption tax is imposed at import, production, where consumption represents the cigarette wholesale and retail levels. Import duty consumption in fiscal year t, price represents has been applied on imported tobacco the weighted average price of cigarettes products since FY2015–6. in fiscal year t. β is the price elasticity of demand of cigarette during FY2011–2 Tonga has significantly increased its and FY2016–7. Using Stata MP (14.0) excise tax on cigarettes in the past few software, we estimated that β=-0.11 (95 years. Excise tax on imported cigarettes percent confidence interval -1.53 to 1.32), rose by 49 percent in FY2016–7 and by an which indicates a 10 percent increase in additional 18 percent in FY2017–8. Excise cigarette prices, will lead to a 1.1 percent tax on locally manufactured cigarettes decrease in cigarette consumption – a increased by 16 percent in FY2016–7 and rather lower sensitivity of cigarette demand by 25 percent in FY2017–8. to the variation of cigarette price. Table 11: Tonga’s tobacco tax structure Excise tax Consumption tax Customs duty Year Imported Locally produced Imported Locally produced Imported Jul.2009- 15% imposed on 15% imposed T$150/1,000 stick T$150/1,000 stick Free Jun.2010 (CIF+excise tax) on excise Jul.2010- 15% imposed on 15% imposed T$200/1,000 stick T$200/1,000 stick Free Jun.2011 (CIF+excise tax) on excise Jul.2011- 15% imposed on 15% imposed T$200/1,000 stick T$200/1,000 stick Free Jun.2012 (CIF+excise tax) on excise Jul.2012- 15% imposed on 15% imposed T$210/1,000 stick T$200/1,000 stick Free Jun.2013 (CIF+excise tax) on excise Jul.2013- 15% imposed on 15% imposed T$250/1,000 stick T$238/1,000 stick Free Jun.2014 (CIF+excise tax) on excise Jul.2014- 15% imposed on 15% imposed T$250/1,000 stick T$238/1,000 stick Free Jun.2015 (CIF+excise tax) on excise 15% imposed on Jul.2015- 15% imposed T$255/1,000 stick T$240/1,000 stick (CIF+Customs duty 15% Jun.2016 on excise +excise tax) 15% imposed on Jul.2016- 15% imposed T$380/1,000 stick T$280/1,000 stick (CIF+Customs duty 15% Jun.2017 on excise +excise tax) 15% imposed on Jul.2017- 15% imposed T$450/1,000 stick T$350/1,000 stick (CIF+Customs duty 15% Jun.2018 on excise +excise tax) Source: Compiled by authors based on tax legislation and confirmed by officers of Ministry of Revenue & Customs 55 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The large increase in excise tax on recommendation for the excise tax to imported cigarettes in July 2016 resulted represent at least 70 percent of the in a 30 percent increase in price of the retail price of tobacco products. most popular imported cigarette brand According to the World Bank team’s analysis in Tonga. The total tax per unit for all of the latest data, the significant tax rise cigarettes (local and imported) rose from in July 2016 raised the proportion of taxes T$292 per 1,000 sticks in FY2014–5 to as a percentage of price of the most sold T$486.63 per 1,000 sticks in FY2017–8. brand of cigarettes to 85 percent, fully compliant with WHO’s recommendation. For imported cigarettes, a customs duty of 15 percent was introduced in Prices of locally produced cigarettes FY2015–6, and excise tax rose from have also risen, though not as much T$250 per 1,000 sticks in FY2014–5 to because of lower excise tax increases T$450 per 1,000 sticks in FY2017–8, applied to them. For locally produced resulting in an increase in total tax cigarettes, excise tax rose from T$238 from T$295.36 per 1,000 sticks to per 1,000 sticks in FY2014–5 to T$350 T$541.71 per 1,000 sticks. The retail per 1,000 sticks in FY2017–8, resulting price of the most popular imported in an increase in total tax from T$273.7 cigarette brand, Pall Mall, sourced from per 1,000 sticks in FY2014–5 to T$402.5 the retail survey in Tonga and from local per 1,000 sticks in FY2017–8. The tobacco agents, increased from T$10 per retail price of the most popular locally pack (20 sticks) in FY2015–6 to T$13 per manufactured brand, Palataisi, sourced pack in FY2016–7. With additional tax from the retail survey and from the local increases, the price rose further to T$15 tobacco agents, increased from T$7.5 per pack (20 sticks) in FY2017–8 (see per pack (20 sticks) in FY2014–5 to Figure 7). T$10 per pack in FY2016-7, and further increased to T$12 per pack in FY2017–8 Tonga leads the way among Pacific (see Figure 8). countries in adhering to WHO’s Figure 7: Tax per unit and retail price of imported cigarettes, T$ Sources: Customs duty, excise tax, consumption tax, and total tax: sourced from the Ministry of Revenue & Customs, Tonga; retail price of FY2014-2015 to FY2016-2017 collected from local tobacco agents; retail price FY2017-2018 reproduced from the retail survey in Tonga. 56 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 12: Price and total tax of a pack of 20 cigarettes (most sold brand) as percentage of retail price, Tonga (compared with selected Pacific countries), 2016 Price of a pack Taxes as % of price of the most sold brand of 20 cigarettes Countries International dollars Value added Specific Ad Import Other (at purchasing tax Total tax Excise Valorem duties taxes power parity) /sales tax Australia 14.86 51.17% 0% 9.09% 0% 0% 60.27% Cook Islands - 63.59% 0% 10.43% 0% 0% 74.02% Fiji 10.31 30.25% 0% 8.26% 0% 0% 38.51% Kiribati 5.69 35.00% 0% 7.03% 0% 0% 42.03% Marshall 2.39 0% 0% 3.85% 40.00% 10.21% 54.06% Islands Micronesia (Federated 2.41 0% 0% 20.00% 40.00% 0% 60.00% States of) Nauru 11.82 0% 0% 0% 50.51% 0% 50.51% New Zealand 14.69 61.90% 0% 3.85% 13.04% 0% 74.94% Niue - 0% 0% 10.75% 58.10% 15.38% 84.23% Palau 6.28 74.07% 0% 0% 0% 0% 74.07% Papua New 9.81 27.78% 0% 9.09% 0% 0% 36.87% Guinea Samoa 5.96 38.54% 0% 13.04% 0% 0% 51.58% Solomon 3.94 19.80% 0% 9.09% 0% 0% 28.89% Islands 67.15% Tonga 7.46 63.33% 0% 1.91% 1.91% 0% (85%)31 Tuvalu 2.95 0% 42.81% 2.03% 5.79% 0.01% 50.63% Vanuatu 6.09 44.44% 0% 6.20% 1.83% 0% 52.48% Source: https://www.who.int/tobacco/global_report/2017/appendix-ix/en/. 31 The latest data from WHO (2016) indicate that total tax represented 67.15% of price of the most sold brand cigarettes in Tonga. According to the World Bank research team analysis, which applied the latest data from FY2016-7, with additional increase in excise tax, the latest total tax represented as high as 85% of price of the most sold brand cigarettes in Tonga. 57 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 8: Tax per unit and retail price of locally manufactured cigarettes, T$ Sources: Excise tax, consumption tax, and total tax: sourced from the Ministry of Revenue & Customs, Tonga; retail price of FY2014-5 to FY2016-7: collected from local tobacco agent; retail price of FY2017-8: reproduced from the retail survey in Tonga. Overall, the large increase in prices Figure 9: Tax revenue, all cigarettes, T$ of cigarettes in FY2016–7 produced a significant fall in consumption of imported and locally manufactured cigarettes, and a slight decrease in overall tax revenues from cigarettes. The total tax per unit levied on cigarettes increased sharply from T$303.99 per 1,000 sticks in FY2015–6 to T$420 per 1,000 sticks in FY2016–7, and to T$486.63 per 1,000 sticks in FY2017–8 (see Figure 9). This resulted in a significant decrease in sales volumes. Between FY2015–6 and FY2016–7, the sales volume dropped Sources: Tax revenue and total tax: sourced from the Ministry of Revenue & Customs, Tonga. sharply by 45 percent (see Figure 10). A further excise tax increase in FY2017–8, saw sales volumes drop by a further 21 Figure 10: Sales volume, all cigarettes, T$ percent. The tax intervention also resulted in a slight decrease in the overall tax revenue from all cigarettes – from T$46.19 million in FY2015–6 to T$35.01 million in FY2016–7, as shown on Figure 9. An additional increase in excise tax in FY2017– 8 also resulted in further decrease in consumption and tax revenues, with tax revenue amounting T$32.03, which represents 3.23 percent of GDP. Nevertheless, more in-depth analysis indicates different outcomes between Sources: Sales volume: sourced from the Ministry of Revenue imported cigarettes and locally & Customs, Tonga; Change rate: calculated by authors. manufactured cigarettes. 58 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga For imported cigarettes, significant Figure 11: Volume and tax revenue of imported cigarettes price increases in FY2016–7 resulted in a large decrease in their consumption, imported volumes, and tax revenues. The approximate 30 percent increase in prices of imported cigarettes (Pall Mall as proxy) resulted in 51.4 percent decrease in imported volume from 122.69 million sticks in FY2015-6 to 59.59 million sticks in FY2016-7, as shown in Figure 11. Since locally manufactured cigarettes Sources: Imported volume, tax revenue, and total tax: sourced were subject to lower excise tax from the Ministry of Revenue & Customs, Tonga. increases than imported cigarettes, their consumption actually increased, Figure 12: Production volume and tax revenue of locally manufactured cigarettes indicating the possibility that some smokers may have shifted from consuming imported cigarettes to cheaper, locally manufactured cigarettes. This also leads to higher tax revenue. The total tax per unit levied on locally manufactured cigarettes increased steadily from T$273.7 per 1,000 sticks in FY2014–5 to T$402.5 per 1,000 sticks in FY2017–8. There was an increase in tax revenue from T$6.37 million to T$10.48 million during the same period (see Figure 12). Sources: Production volume, tax revenue, and total tax: sourced from the Ministry of Revenue & Customs, Tonga. Cigarettes have become less affordable in Tonga following the country’s Figure 13: Cigarette affordability, 2013–2017 significant 2016 excise tax rise. There are two indexes measuring cigarette affordability. The Income Purchasing Capacity (IPC) defines cigarette affordability as GDP per capita/price – the higher the IPC, the more affordable cigarettes are, and vice versa. The other is the retail index price (RIP), which defines cigarette affordability as the percentage of per capita GDP required to buy 100 packs of cigarettes. The higher the RIP, the less affordable cigarettes are, and vice versa. As shown in Figure 13, Sources: IPC and RIP: calculated from GDP per capita affordability of cigarettes in Tonga measured (sourced from Word Bank Open Data) and retail price (reproduced from local tobacco agent and by these two indexes has declined as a the retail survey in Tonga) result of increased excise tax and price. 59 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Large tobacco tax rises are powerful to overtake Pall Mall to become the most enough to change the market share popular cigarette brand in Tonga. of the most popular cigarette brand in Tonga. Prior to the large excise tax Price, not health, is the main driver increase on imported cigarettes in for smokers changing their tobacco FY2016–7, Pall Mall had been the most consumption behaviors. Among smokers consumed cigarette brand in Tonga, with who have decreased their consumption, a market share of over 50 percent in over 70 percent indicated that price is the FY2014–5 and FY2015–6. However, the main reason for the changed behaviors. large excise tax increase on imported Only around 10 percent indicated that cigarettes in FY2016–7 led to a price hike they changed behaviors because smoking for imported cigarettes, and a significant was unhealthy (see Figure 15). This number of smokers simply switched to indicates that price interventions provide cheaper, locally manufactured cigarettes. one of the most effective tools for tobacco As shown in Figure 14, tax increases consumption behavior change that countries imposed on imported cigarettes in July 2016 should be applying, in addition to health allowed Palataisi, a locally manufactured promotion campaigns and activities, cigarette brand with less tax imposed on it, tobacco cessation programs, and others. Figure 14: Market share of different cigarette brands in Tonga, 2011–2017 Sources: Ministry of Revenue & Customs, Tonga. Figure 15: Reasons for behavior change Source: Baseline and endline survey data (2017). 60 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The majority of smokers were aware of Only an additional 2 percent of cigarette the tobacco tax increase, and knowledge smokers shifted to Tapaka Tonga by the about the tax has increased over time. endline survey (see Figure 16). According to the baseline household survey, around 62 percent of tobacco Figure 16: Behavior change following FY2016–7 (baseline) tobacco tax increase and FY2017–8 (endline) users were aware of the increase in excise tobacco tax increase tax imposed in July 2016. Nevertheless, several of them suggested that they only learned about the price increase when they visited shops to purchase cigarettes. This recommendation for improving communication efforts regarding the tax was shared with the government by the research team. It is important to note that the knowledge about tobacco tax has increased over time. According to the endline survey, approximately 85 percent of smokers were aware of the July 2017 tobacco tax increase. This is likely a result of better communication efforts by the Source: Baseline and endline survey data (2017). government, as well as the fact that it takes time for knowledge to reach all The increase in excise tax on tobacco communities. in FY2016–7 has greater impacts on the less well-off. While approximately 18 The large tobacco excise tax increase in percent of overall smokers decreased FY2016–7 led 18.4 percent of smokers consumption of cigarettes, there is greater to decrease their cigarette consumption, impact among smokers from lower socio- but 20 percent of smokers shifted economic status. As shown in Figure 17, from cigarettes to Tapaka Tonga. among smokers who are less well-off According to the household survey, (quintile 1 and 2), over 21 percent approximately 18 percent of smokers decreased consumption of cigarettes. decreased consumption of cigarettes, There is less impact among the wealthy which indicates that the price intervention group. Only 13.9 percent of quintile 5 had a significant impact. However, 20 decreased consumption of cigarettes percent of cigarette smokers have shifted and nearly 60 percent of them did not to Tapaka Tonga, which is dry loose change behaviors, which may indicate tobacco leaves for roll-your-own. Tapaka that the wealthiest group is less sensitive Tonga is cheap, widely available in shops, to price changes. Nevertheless, it is and is not subject to excise tax. Nevertheless, important to note that Tapaka Tonga the endline survey has shown that a has become the major substitute for further tobacco tax increase in FY2017–8 manufactured cigarettes across all socio- prompted an additional 20 percent of economic status groups. smokers to decrease smoking. The average number of manufactured cigarette sticks consumed by smokers per day was 16 during the baseline. This was reduced to only 8.8 sticks by the endline survey. The data also show that this substantial reduction is not significantly accounted for by switching to Tapaka Tonga. 61 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 17: Behavioral responses on cigarette There was a decrease in the number consumption by wealth quintiles after the FY 2016/7 increase in excise tax of daily smokers of manufactured cigarettes. However, most Tapaka Tonga users smoked on a daily basis. During the baseline survey, 91.15 percent of manufactured cigarette smokers smoked daily. However, by the endline survey, only 83 percent of them smoked daily (see Figure 18). It appears that some of these formerly daily smokers now smoked less frequently, and may have shifted to smoke a few times a week or less. However, it is important to note that 92.6 percent of Tapaka Tonga smokers Source: Endline survey data (2017). smoked on a daily basis (Figure 19).32 Tapaka Tonga is cheap and widely Figure 18: Proportion of manufactured cigarette smokers available in shops. The shift among smokers from consuming manufactured cigarettes to consuming Tapaka Tonga is alarming, and will most likely dilute any potential positive effects on population health as a result of the NCD tax policy. Tapaka Tonga is currently not subject to excise tax, and is cheap and widely available in shops countrywide. Hence, it is crucial for the Government of Tonga to consider introducing fiscal interventions and other non-tax interventions to address Source: Baseline and endline survey data (2017). this issue. A significant part of the Tongan population also mistakenly thought that Tapaka Tonga is not harmful (or is less Figure 19: Proportion of Tapaka Tonga smokers harmful) to health than manufactured cigarettes. Moreover, it is common for Tapaka Tonga packaging to be labeled ‘organic’, which can be misleading, as shown in Picture 3. Tapaka Tonga has become increasingly affordable compared to manufactured cigarettes. While the price of manufactured cigarettes increased significantly as a result of the tobacco tax intervention, the price of Tapaka Tonga stayed mostly the same Source: Endline survey data (2017). (see Figure 20). After July 2017 it rose slightly from T$5 to T$6 – largely due to higher demand. 31 The World Bank research team became aware of the extent of Tapaka Tonga use during the analysis of the baseline household surveys, and hence more ques- tions to understand the behaviors of Tapaka Tonga users were added for the endline survey. 62 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Picture 3: Tapaka Tonga as sold in shops Figure 21: Spending on manufactured cigarettes in the past week, T$ Source: Endline survey data (2017). Figure 22: Spending on Tapaka Tonga, T$ Figure 20: Retail prices of Tapaka Tonga and manufactured cigarettes over time Source: Endline survey data (2017). There are now more Tapaka Tonga users than manufactured cigarette users. There was very limited information on the proportion and behaviors of Tapaka Tonga users in the country before this study Source: Baseline and endline survey data (2017). was conducted. Nevertheless, according Tapaka Tonga smokers spent much to the STEPS survey 2012, 85.25 percent of less than manufactured cigarette daily smokers in Tonga smoked manufactured smokers. The majority of manufactured cigarettes, which may indicate that daily cigarette smokers spent T$11–20 per Tapaka Tonga smokers were the minority at week, while over one-third spent as much that time. However, according to the endline as T$21–50 per week. However, the survey conducted in late 2017, 51.38 percent majority of Tapaka Tonga smokers spent of Tongan smokers smoked Tapaka Tonga, only T$1–10 per week, indicating how while 37.27 percent of smokers consumed cheap and affordable Tapaka Tonga is manufactured cigarettes (see Figure 23). (see Figure 21 and Figure 22). Around 11.34 percent consumed both Tapaka Tonga and manufactured cigarettes. 63 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 23: Type of smokers in Tonga (November 2017) Food Obesity and poor diet are the most significant risk factors contributing to the country’s high NCD burden, and the government has prioritized food tax policy as a key way to address it. Tonga has pioneered the use of food tax policy in the Pacific to address demand for a number of products perceived as unhealthy, imposing excise tax and/or import duties designed to reduce the consumption of fatty foods (e.g. turkey tails, mutton flaps, chicken leg quarters), sugary foods (e.g. ice cream, chocolates, biscuits), sugar-sweetened beverages (SSB), and instant noodles. At the same Source: Endline survey data (2017). time the government has abolished 15 percent consumption tax on imported In summary, Tonga has achieved much fruits and vegetables perceived as healthy, through its continuing tobacco tax and in anticipation that the price of these price increases. Tax as a percentage products will fall and their consumption of retail price is 85 percent, among the will rise. highest in world. The significant increase of excise and price, particularly in FY2016 While the food tax policy covers many –7 and FY2017–8, reduced cigarette food products, the study team focused consumption dramatically. Tonga is also on just a few key products consumed one of very few countries to see reduced by the majority of Tongans. Among the revenue along with tax hikes – the result food products subject to NCD tax, the study of flaws in the tobacco tax structure, which team selected turkey tails, mutton flaps, include: (i) differential excise rates on chicken leg quarters, ice cream, SSB, and imported and locally produced cigarettes instant noodles. Among the imported food that enlarged the price gap between products subject to the 15 percent tax imported brands and local brands, and exemption, the study team chose apples therefore encouraged smokers to switch and oranges. The detailed tax structure to local brands when tax rose; (ii) the of selected food and drinks is listed in absence of tax on Tapaka Tonga, which Table 13. has hampered government efforts to use tax and price measures to curb tobacco Prices have increased following the use. In order to achieve a win-win (less imposition of excise tax/import duty consumption of tobacco products and across all products. For the excise tax more revenue), it is necessary for Tonga to affect consumption behaviors, retail to apply a unified excise tax rate on prices must be changed. According to the imported and locally manufactured retail survey and consumer price index, cigarette brands, and Tapaka Tonga. retailers have responded to the imposition of excise, and prices have increased across all products (see Figure 24). 64 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 13: Tax structure of food and SSBs (July 2014–June 2018) Excise tax Consumption tax Custom duty Product Year Locally Locally Imported Imported Imported produced produced Jul.2014- 15% imposed on Free NA NA Free Jun.2015 (CIF+Customs Duty) Jul.2015- 15% imposed on Free NA NA 15% Jun.2016 (CIF+Customs Duty) Turkey 15% imposed on Jul.2016- T$1.50/kg NA (CIF + Customs duty NA 15% Jun.2017 + excise tax) 15% imposed on Jul.2017- T$1.50/kg NA (CIF + Customs duty NA 15% Jun.2018 + excise tax) FY2014- 15% imposed on Free NA NA 15% 2015 (CIF+customs duty) FY2015- 15% imposed on Free NA NA 15% 2016 (CIF+customs duty)) Ice cream 15% imposed on FY2016- T$1.5/liter NA (CIF+customs duty NA 15% 2017 +excise tax) 15% imposed on FY2017- T$1.5/liter 25 seniti/liter (CIF+customs duty NA 15% 2018 +excise tax) Jul.2014- 15% imposed on Free NA NA Free Jun.2015 (CIF+customs duty) Jul.2015- 15% imposed on Free NA NA Free Jun.2016 (CIF+customs duty) Mutton flaps Jul.2016- 15% imposed on Free NA NA 15% Jun.2017 (CIF+customs duty) Jul.2017- 15% imposed on T$1.15/kg NA NA 15% Jun.2018 (CIF+customs duty) Jul.2014- 15% imposed on Free NA NA Free Jun.2015 (CIF+customs duty) Jul.2015- 15% imposed on Free NA NA Free Chicken leg Jun.2016 (CIF+customs duty) quarters Jul.2016- 40 seniti per 15% imposed on NA NA Free Jun.2017 kg (CIF+excise) Jul.2017- 40 seniti per 15% imposed on NA NA Free Jun.2018 kg (CIF+excise) Jul.2014- 15% imposed on Free NA NA 15% Jun.2015 (CIF+customs duty) Jul.2015- 15% imposed on T$1/kg NA NA Free Instant Jun.2016 (CIF+customs duty) noodles Jul.2016- 15% imposed on T$1/kg NA NA Free Jun.2017 (CIF+excise tax) Jul.2017- 15% imposed on T$2/kg NA NA Free Jun.2018 (CIF+excise tax) Jul.2014- 15% imposed on NA NA NA 15% Jun.2015 (CIF+customs duty) Jul.2015- 15% imposed on NA NA NA 15% Jun.2016 (CIF+customs duty) Fruits Jul.2016- NA NA Free NA Free Jun.2017 Jul.2017- NA NA Free NA Free Jun.2018 65 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 13: Tax structure of food and SSBS (July 2014–June 2018) (continued) Excise tax Consumption tax Custom duty Product Year Locally Locally Imported Imported Imported produced produced Jul.2012- 15% imposed on Free NA NA 15% Jun.2013 (CIF+customs duty) Jul.2013- 15% imposed on .50/liter NA NA 0% Jun.2014 (CIF+customs duty) Jul.2014- 15% imposed on .50/liter NA NA 0% Jun.2015 (CIF+customs duty) Jul.2015- 15% imposed on .50/liter NA NA 0% Jun.2016 (CIF+customs duty) Jul.2016- 15% imposed on 1/liter NA NA 0% Jun.2017 (CIF+customs duty) SSB Sugar content not exceeding 15% imposed on 0 NA NA 15% 5 gram per (CIF+excise Tax) 100ml Sugar content Jul 2017- not exceeding 15% imposed on T$1.50/liter NA NA Free Jun 2018 20 gram per (CIF+excise Tax) 100ml Sugar content exceeding 15% imposed on T$4/liter NA NA Free 20 gram per (CIF+excise Tax) 100ml Source: Compiled by authors based on tax legislation and confirmed by officers of Ministry of Revenue & Customs Figure 24: Price rises after the excise tax/import duty increase 66 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga While significant excise tax was Figure 25: Awareness of the excise tax increase in July 2016 on selected food products as of June 2017 imposed on a number of food products in July 2017, less than 20 percent of the population were aware of its implementation. Over half of the population were not aware of the implementation of the food tax policy, while over a quarter only became aware of it when buying food at stores (see Figure 25). From focus group discussions, many households reported that they thought it was the Chinese shops that voluntarily increased the price to increase their profits. Among those who were aware of the food tax, most were aware of the excise Source: Baseline survey data (2017). tax on chicken leg quarters and mutton flaps. Chicken leg quarters and mutton Figure 26: Knowledge about food products subjected to tax increase in July 2016 flaps are common meat consumed in Tonga, and it is understandable that more consumers were aware of the food tax being imposed on these products (see Figure 26). Substantially fewer Tongans were aware that there was an excise tax being imposed on other food products, including SSB and imported instant noodles, which were also popular products. Households in the “less well-off” category were more aware than their wealthier counterparts of the price increases on selected food products Source: Baseline survey data (2017). (not necessarily knowing that it was because of the tax), indicating this group’s greater price sensitivity. Figure 27: Notices about price changes of selected products by wealth quintiles From the household survey, more households in the less well-off category (wealth quintile 1 and 2) noticed the price increase in chicken leg quarters, mutton flaps, and turkey tails. Nevertheless, fewer of the least well-off households (wealth quintile 1) noticed the price change of ice cream, which could indicate that ice cream may not be a common food item among this particular group (see Figure 27). Source: Baseline survey data (2017). 67 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Mutton flaps Mutton flaps are considered the fattiest part of lamb, and are one of the most popular dishes in Tonga. Mutton flaps are not produced domestically in Tonga, but are imported. Pre-cooked mutton flaps are shown in Picture 4. Mutton flaps cooked as Tongan dishes are shown in Picture 5. A customs duty of 15 percent was imposed on mutton flaps in FY2016–7 and an excise tax of T$1.15 per kg Picture 4: (left) Pre-cooked mutton flaps Picture 5: (right) Cooked mutton flaps as sold in shops was introduced in FY2017–8, resulting in a significant increase in retail price. The total tax per unit levied on mutton from 863,940 kg to 1,043,970 kg. flaps increased from T$0.98 per kg in However, once the 15 percent import duty FY2014–5 to T$4.24 per kg in FY2017–8 was imposed in FY2016–7, the import (see Figure 28). The average retail price of volume of mutton flaps dropped by about mutton flaps, according to the consumer 35 percent, and dropped again following price index and retail survey, increased the introduction of an excise tax of T$1.15 from T$10.50 per kg in FY2015–6 to per kg in FY2017–8, which generated a T$15.38 per kg in FY2017–8. total tax of T$4.24 per kg. It is important to note that despite lower import volumes, Import volumes of mutton flaps tax revenues from import duty and excise decreased following the import duty tax have increased by 90 percent in and excise tax. A drop in total tax in FY2016–7 and a further 59 percent FY2015–6 drove up the imported volume increase in FY2017–8 (see Figure 29). Figure 28: Tax per unit and retail price of mutton flaps Sources: Customs duty, excise tax, consumption tax, and total tax: sourced from the Ministry of Revenue & Customs, Tonga; Retail price: sourced from the Ministry of Finance and National Planning, Tonga. 68 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 29: Consumption, tax revenue and total tax per unit of mutton flaps Sources: Imported volume, tax revenue, and total tax: sourced from the Ministry of Revenue & Customs, Tonga. Figure 30: Change in Figure 31: Change in Approximately half of Tongans changed consumption of mutton consumption of mutton their consumption of mutton flaps flaps (baseline) flaps (endline) after the tax policy (Figure 30). This number increased to 55.24 percent when the price further increased as a result of imposition of excise tax in July 2017 (see Figure 31). Over 78 percent of those who changed behaviors after the July 2016 tax reduced their consumption of mutton Source: Baseline survey data Source: Endline survey data (2017). (2017). flaps, but gradually, more people shifted to alternative products. When the import duty was imposed in July 2016, most Figure 32: Consumption behavior change following the people reduced their consumption of imposition of tax on mutton flaps mutton flaps, while around 16 percent opted for alternative products. Following the July 2017 tax increase, as much as 40 percent of the population no longer simply reduced consumption of mutton flaps but went for alternative products (see Figure 32). Salted beef is the main substitute for mutton flaps. Source: Baseline and endline survey data (2017). 69 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Price is the main factor for the changed Figure 33: Reason for changing consumption behaviors of mutton flaps consumption behaviors of mutton flaps. From both baseline and endline survey, 98–99 percent of survey respondents cited price as the main reason for changing consumption behaviors of mutton flaps. Nevertheless, an increasing number of people during the endline survey cited health as a main reason for behavior change (see Figure 33). Tax has affected the behaviors of those who frequently consumed mutton flaps. Source: Baseline and endline survey data (2017). Following the imposition of import duty Figure 34: Frequency of consumption of mutton flaps in FY2016–7, only 1.79 percent of the (baseline and endline surveys) population consumed mutton flaps five to six times a week, and just over 7 percent consumed them three to four times a week. The number of frequent consumers of mutton flaps fell further following the increased tax in FY2017–8. The percentage of people not consuming mutton flaps rose from 19 percent to 24 percent during the same period. Nevertheless, about 50 percent of Tongans still consumed mutton flaps one or two times per week (see Figure 34). Source: Baseline and endline survey data (2017). Turkey tails Turkey tails are the fattiest part of turkey, and are one of the most popular dishes in Tonga. Turkey tails are not produced domestically in Tonga, and hence all of them in Tongan markets are imported. Pre-cooked turkey tails are shown in Picture 6. Cooked turkey tails are shown in Picture 7. An excise tax of T$1.5 per kg of turkey tails was introduced in FY2016–7, resulting in a price rise of approximately 12 percent. Prior to the introduction of excise tax, the government imposed an import duty in FY2015–6. This led to Picture 6: (above) Pre-cooked turkey tails an increase in total tax from T$1.03 Picture 7: (below) Cooked turkey tails per kg in FY2015–6 to T$2.31 per kg in 70 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga FY2016–7. According to the retail survey Figure 35: Tax per unit and retail price of turkey tails and consumer price index, the retail price of turkey tails increased from T$5.27 per kg in FY2014–5 to T$6.2 per kg in FY2017–8 (see Figure 35). The largest increase in price, by 12 percentage points, was observed in FY2016–7 when the significant excise tax was introduced. Imports of turkey tails (a proxy for consumption) had fallen significantly since the import duty was imposed in FY2015–6, accompanied by a continuous rise in both price and tax revenue. The total tax per unit levied on turkey Sources: Customs duty, excise tax, consumption tax, and total tax: sourced from the Ministry of Revenue & Customs, Tonga; tails increased steadily from T$0.49 Retail price of FY2014-5, FY2015-6, and FY2017-8: sourced per kg in FY2014–5 to T$2.68 per kg from the Ministry of Finance and National Planning, Tonga; Retail price of FY2016-7: reproduced from the retail survey in FY2017–8, resulting in a decline in in Tonga. import volume from 412,390 kg to 147,720 Figure 36: Consumption, tax revenue and kg, and an increase in tax revenue, from total tax per unit of turkey tails T$201,430 to T$396,620 (see Figure 36). Nearly half of Tongans changed their consumption behaviors of turkey tails following the July 2016 tax increase, with 82 percent of them consuming less turkey tails. The number of people who changed behaviors increased by 10 percentage points by July 2017. Approximately 6–7 percent of Tongans stopped eating turkey tails following the tax intervention. Chicken is the main substitute for turkey tails. (see Figure 37, Figure 38, and Figure 39). Sources: Imported volume, tax revenue, and total tax: sourced from the Ministry of Revenue & Customs, Tonga. Figure 37: Changed consumption behaviors Figure 38: Changed consumption behaviors of turkey tails at baseline of turkey tails at endline Source: Baseline survey data (2017). Source: Endline survey data (2017). 71 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 39: Behaviors following the imposition Figure 40: Reasons for changing consumption of tax on turkey tails behaviors of turkey tails Source: Baseline and endline survey data (2017). Source: Baseline and endline survey data (2017). Price is the main reason for changing Figure 41: Tax per unit and retail price of chicken legs turkey tails consumption behaviors. Over 90 percent of respondents cited that price was the main reason for changing turkey tails consumption behaviors (see Figure 40). While only a small minority cited health as a reason for changing consumption after the July 2016 tax intervention, this proportion increased to nearly 50 percent after the July 2017 tax intervention. This may indicate that the health awareness has gradually increased. Chicken leg quarters Sources: Customs duty, excise tax, consumption tax, An excise tax of T$0.40 per kg was and total tax: sourced from the Ministry of Revenue introduced in FY2016–7, resulting in a & Customs, Tonga; Retail price: sourced from the Ministry of Finance and National Planning, Tonga. total tax increase from T$0.24 per kg to T$0.76 per kg, and a significant price increase. No customs duty was levied on While the price of chicken leg quarters chicken legs from FY2014–5 to FY2016– increased by 27 percent following 7. The average retail price of chicken legs, the imposition of excise tax, import according to the consumer price index volumes only slightly decreased. The and retail survey, increased by 27 percent, import volumes of chicken leg quarters from T$2.95 per kg in FY2015–6 to T$3.75 decreased from 8,965,000 kg to 8,011,620 per kg in FY2016–7 (see Figure 41). kg in FY2016–7. However, the revenue generated by the excise tax grew by 178 percent, from T$2.175 million to T$6.055 million during this period (see Figure 42). 72 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga In contrast to mutton flaps and turkey Figure 42: Consumption, tax revenue and total tax per unit of chicken legs tails, most people still consume chicken regularly, despite the price increase. While most Tongans reduced consumption of mutton flaps and turkey tails following the tax intervention and price increase, over 70 percent of Tongans did not change their chicken leg quarters consumption behaviors, as shown in Figure 43. Most people still consumed chicken leg quarters three to six times a week, as shown in Figure 44. Among the minority who changed behaviors, most chose tinned fish as substitute, but some chose less healthy products like sausages, corned Sources: Imported volume, tax revenue, and total tax: sourced beef and salted beef. from the Ministry of Revenue & Customs, Tonga. Figure 43: Changed consumption behaviors of Ice cream chicken leg quarters at baseline An excise tax of T$1.5 per liter was introduced in FY2016–7, resulting in an increase in total tax from T$0.98 per liter in FY2015–6 to T$2.69 per liter in FY2016–7, and a significant price increase. The average retail price of ice cream from the retail survey and consumer price index increased from T$11.61 per bowl in FY2014–5 to T$15.5 per bowl in FY2017–8. The price increased by over 20 percentage points after the introduction of excise tax in FY2016–7, from T$12.73 to T$15.36 (Figure 45). Source: Baseline survey data (2017). The imposition of excise tax in FY2016 Figure 44: Frequency of consumption of chicken leg –7 led to a decrease in imported quarters (baseline and endline surveys) ice cream and also contributed to government tax revenues. The increase in excise tax on ice cream in FY2016–7 resulted in significant price increases and a significant reduction in the import volumes of ice cream. It also contributed to a significant increase in tax revenue, from T$1,583,530 to T$3,113,080. It is important to note that import volumes continued to decline in FY2017–8, indicating the effectiveness of the policy. Although this further decline in imported volume in FY2017–8 resulted in a decrease in tax revenue compared to the year before, the amount of revenue collected in FY2017–8 Source: Baseline and Endline survey data (2017). 73 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga was still significantly higher than that of Figure 47 and Figure 48). Of those who FY2015–6 (before the introduction of the changed behaviors, 90 percent reduced excise tax), as per Figure 46. their consumption of ice cream. This reduction was observed particularly Approximately 40 percent of ice cream among those who consumed ice cream consumers changed their behaviors very frequently (three to six times per in July 2016, and this proportion week), as per Figure 49. increased to 51 percent in July 2017 (see Figure 45: Tax per unit and retail price of ice cream Figure 46: Consumption, tax revenue and total tax per unit of ice cream Sources: Customs duty, excise tax, consumption tax, and total Sources: Imported volume, tax revenue, and total tax: sourced tax: sourced from the Ministry of Revenue & Customs, Tonga; from the Ministry of Revenue & Customs, Tonga. Retail price: sourced from the Ministry of Finance and National Planning, Tonga. Figure 47: Changed Figure 48: Changed Figure 49: Frequency of consumption of ice cream consumption behaviors consumption behaviors of (baseline and endline surveys) of ice cream at baseline ice cream at endline Source: Baseline survey data Source: Endline survey data (2017). (2017). Source: Baseline and endline survey data (2017). 74 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Imported instant noodles Figure 50: Tax per unit and retail price of imported instant noodles Customs duty has not applied to imported instant noodles since FY2015 –6, but an excise tax of T$1 per kg was introduced in FY2015–6, rising to T$2 per kg in FY2017–8. The retail price of instant noodles, for example the “Supreme” brand from the retail survey, increased from T$0.7 per pack in FY2016–7 to T$1 per pack in FY2017–8. The average retail price of instant noodles (two-minute maggi), provided by the Ministry of Finance and National Planning of Tonga increased from T$0.88 per pack in FY2014–5 to T$1.14 per pack in FY2017–8. The total Sources: Customs duty, excise tax, consumption tax, and total tax: sourced from the Ministry of Revenue & Customs, Tonga; tax per unit levied on instant noodles Retail price: sourced from the Ministry of Finance and National increased from T$0.88 per kg in FY2014– Planning, Tonga. 5 to T$3.14 per kg in FY2017–8 (see Figure 51: Consumption, tax revenue and Figure 50). total tax per unit of imported instant noodles While the imposition of excise tax in FY2015–6 initially led to a significant decline in the import of instant noodles, imports during FY2017–8 were nearly the same as during FY2015–6. The increase in excise tax on imported instant noodles in FY2015–6 resulted in a sharp decrease in imports from 2,083,410 kg to 885,280 kg, and a further decline continued in FY2016–7. However, the imported volume rebounded from 438,290 kg in FY2016–7 to 805,540 kg in FY2017–8, even though the excise tax was doubled Sources: Imported volume, tax revenue, and total tax: sourced to T$2 per kg in that year (see Figure 51). from the Ministry of Revenue & Customs, Tonga. Figure 52: Notice of price Figure 53: Changed While two-thirds of Tongans were change of imported instant consumption behaviors of aware of the price increase of imported noodles at baseline imported instant noodles at baseline instant noodles, few people changed their consumption behaviors following the tax interventions. According to the household survey, over 66 percent of the population were aware of the price increase of imported instant noodles. However, less than 30 percent of the population changed or reduced consumption of imported instant noodles, following the tax intervention (see Figure 52 and Figure Source: Baseline survey data Source: Baseline survey data 53). The low price base and small changes (2017). (2017). in prices may be the key factor. 75 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Sugar-sweetened beverages An excise tax of T$0.5 per liter per liter in FY2012–3 to T$1.85 per liter was introduced on sugar-sweetened in FY2017–8 resulted in an increase beverages (SSB) while customs duty in tax revenue from T$1,543,000 to was abolished in FY2013–4; excise tax T$8,413,000. While the import volume was increased to T$1 in FY2016–7, and dropped in FY2016–7, it increased a new tax formula based on sugar significantly in FY2017–8 (see Figure 55). content was introduced in FY2017–8. This resulted in an increase in total tax While over three-quarters of Tongans from T$0.37 per liter in FY2012–3 to were aware of the price increase of T$0.73 per liter in FY2013–4, and to SSB, far less than half of the population T$1.85 per liter in FY2017–8. The retail changed their consumption behaviors price of SSB, reported from the retail survey, following the tax interventions. for example Coca-Cola, increased from According to the household survey, nearly T$2 per can in FY2015–6 to T$2.4 per can 76 percent of the population were aware in FY2016–7, and further to T$2.5 per can of the price increase of SSB (see Figure in FY2017–8 (Figure 54). 56), but just 38 percent of the population changed and reduced their consumption While there was a 20 percent price (see Figure 57). It is important to note increase in FY2016–7, the imported that over one-third of Tongans do not volume of SSB dropped slightly, but consume SSB. Among SSB consumers, rebounded in FY2017–8. The steady most consumed them once or twice per increase in SSB tax rates from T$0.37 week (see Figure 58). Figure 54: Tax per unit and retail price of sugar-sweetened beverages Sources: Customs duty, excise tax, consumption tax, and total tax: sourced from the Ministry of Revenue & Customs, Tonga; Retail price: reproduced from the retail survey in Tonga. 76 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 55: Consumption, tax revenue and total tax per unit of SSBs Sources: Imported volume, tax revenue, and total tax: sourced from the Ministry of Revenue & Customs, Tonga. Figure 56: Notice of price change of Figure 57: Changed consumption Figure 58: Frequency of consumption SSB at baseline behaviors of SSB at baseline of SSBs (baseline and endline surveys) Source: Baseline survey data (2017). Source: Baseline survey data (2017). Source: Baseline and endline survey data (2017). 77 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Alcohol Alcohol in Tonga is subject to excise locally manufactured beers and spirits tax, consumption tax and import duty, have lower excise rates than their imported with the aim of reducing harmful alcohol counterparts. Customs duty was imposed consumption. Beers and spirits are on imported alcohol in FY2015–6 (see subject to two different excise rates, while Table 14). Table 14: Tonga alcohol tax structure Excise tax Consumption tax Customs duty Year Locally Locally Imported Imported Imported produced produced Jul.2014- 15% imposed on 15% imposed T$50/Lal T$10/Lal Free Jun.2015 (CIF+excise tax) on excise 15% imposed on Jul.2015- 15% imposed T$50/Lal T$15/Lal (CIF+Customs duty 15% Jun.2016 on excise +excise tax) 15% imposed on Jul.2016- 15% imposed T$50/Lal T$15/Lal (CIF+Customs duty 15% Jun.2017 on excise +excise tax) 15% imposed on Jul.2017-Jun. 15% imposed T$60/Lal T$20/Lal (CIF+Customs duty 15% 2018 on excise +excise tax) Tongan spirits tax structure (whisky, rum, gin, vodka, liqueurs) Jul.2014- 15% imposed on 15% imposed T$42/Lal T$21/Lal Free Jun.2015 (CIF+excise tax) on excise 15% imposed on Jul.2015- 15% imposed T$50/Lal T$25/Lal (CIF+Customs duty 15% Jun.2016 on excise +excise tax) 15% imposed on Jul.2016- 15% imposed T$50/Lal T$25/Lal (CIF+Customs duty 15% Jun.2017 on excise +excise tax) 15% imposed on Jul.2017-Jun. 15% imposed T$60/Lal T$30/Lal (CIF+Customs duty 15% 2018 on excise +excise tax) *Lal: Liters of alcohol contained in a mixture Source: Compiled by authors based on tax legislation and confirmed by officers of Ministry of Customs and Revenue. Beer An increase in excise tax on imported For locally manufactured beer, excise beer saw a rise from T$50/Lal in tax increased from T$10/Lal in FY2014–5 to T$60/Lal in FY2017–8, with FY2014–5 to T$15/Lal in FY2015–6, a customs duty of 15 percent being and further increased to T$20/Lal in introduced in FY2015–6, leading to FY2017–8. This resulted in an increase in price increases. This resulted in an total tax from T$11.5 per Lal in FY2014–5 increase of total tax from T$66.39 per Lal to T$23 per Lal in FY2017–8 (Figure in FY2014–5 to T$83 per Lal in FY2017–8 60). The retail price of a bottle of (see Figure 59). The retail price of beer, local beer, according to the retail taking Heineken as an example from survey, increased from T$2.50 to the retail survey, increased from T$3 per T$2.80, which is 12 percent increase in bottle in FY2016–7 to T$3.5 per bottle price. in FY2017–8, which is approximately 16 percent increase in price. 78 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 59: Tax per unit of imported beer Figure 60: Tax per unit of local beer Sources: Customs duty, excise tax, consumption tax, and total Sources: Excise tax, consumption tax, and total tax: sourced tax: sourced from the Ministry of Revenue & Customs, Tonga; from the Ministry of Revenue & Customs, Tonga; Retail price: Retail price: reproduced from the retail survey in Tonga. reproduced from the retail survey in Tonga. The total tax per unit of imported and Figure 61: Tax per unit of beer locally manufactured beer increased from T$63.82 per Lal in FY2014–5 to T$80.39 per Lal in FY2017–8 (see Figure 61). The total tax per unit levied on imported beer increased steadily from T$66.39 per Lal in FY2014–5 to T$83 per Lal in FY2017–8, resulting in a decrease in imported volume and an increase in tax revenue in FY2017–8 from T$5,547,600 to T$6,626,400. The import volume in FY2016–7 also reduced from 98,000 Lal to 79,830 Lal, as per Figure 62. Sources: Customs duty, excise tax, consumption tax, and total tax: sourced from the Ministry of Revenue & Customs, Tonga. The total tax per unit levied on local beer Figure 62: Consumption, tax revenue increased steadily from T$11.5 per Lal in and total tax per unit of imported beer FY2014–5 to T$23 per Lal in FY2017–8, resulting in an increase in tax revenue, from T$472,000 to T$837,000. Comparing FY2014–5 and FY2017–8, the production volume decreased slightly from 4,100 Lal to 3,640 Lal (see Figure 63). The volume of locally produced beer is significantly smaller than that of imported beer, and has fluctuated in recent years due to disruptions in the production and manufacturing process. Sources: Imported volume, tax revenue, and total tax: sourced from the Ministry of Revenue & Customs, Tonga. 79 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 63: Consumption and tax revenue, Figure 64: Consumption, tax revenue locally manufactured beer and total tax per unit of beer Sources: Producing volume, tax revenue, and total tax: Sources: Sales volume, tax revenue, and total tax: sourced sourced from the Ministry of Revenue & Customs, Tonga. from the Ministry of Revenue & Customs, Tonga. The total tax per unit levied on beer Figure 65: Changes in beer consumption frequency increased steadily from T$63.82 per Lal in FY2014–5 to T$80.39 per Lal in FY2017–8, resulting in drop in sales. However, tax revenue from alcohol sales increased (see Figure 64). Higher beer prices meant people consumed beer less frequently. For example there were fewer alcohol drinkers consuming it five to six times per week and three to four times per week following the FY2017–8 alcohol tax increase, and there was a 10 percentage point increase among those who drink alcohol once per Source: Baseline and endline survey data (2017). month (see Figure 65). Figure 66: Weekly beer consumption People also consumed less beer overall. Prior to the FY2017–8 tax increase, 34 percent of beer drinkers consumed more than 15 standard drinks per week. This proportion reduced to 24.9 percent following the tax increase (see Figure 66). It appears that this group may have shifted to drinking less than 10 standard drinks per week, as there was a proportionate increase among this group following the FY2017–8 tax intervention. Source: Baseline and endline survey data (2017). 80 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 67: Change in beer consumption behaviors Overall, 30 percent of beer drinkers consumed less. The main reason for this behavior change was price (see Figure 67, Figure 68). Nevertheless, it is important to note that 67 percent of beer drinkers did not change behaviors during this period, and continued drinking. Spirits A customs duty of 15 percent was introduced for imported spirits in FY2015–6. Excise tax increased from T$42/Lal in FY2014–5 to T$50/Lal in FY2015–6, and further increased to T$60/Lal in FY2017–8. This resulted in an Source: Endline survey data (2017). increase in total tax from T$55.38 per Figure 68: Reasons for changing alcohol Lal in FY2014–5 to T$87.75 per Lal in consumption behaviors FY2017–8 (see Figure 69). For locally produced spirits, excise tax increased from T$21/Lal in FY2014–5 to T$25/Lal in FY2015–6, then increased to T$30/Lal in FY2017–8. This resulted in an increase in total tax from T$24.15 per Lal in FY2014–5 to T$34.5 per Lal in FY2017–8, (see Figure 70). The retail price of alcohol, taking Spirit Crown as an example from the retail survey, increased from T$25 per bottle in FY 2016–7 to T$26 per bottle in FY2017–8. In total, the total tax per unit of imported Source: Endline survey data (2017). and local manufactured spirits increased Figure 69: Tax per unit of imported spirits from T$52.54 per Lal in FY2014–5 to T$79.62 per Lal in FY2017–8 (see Figure 71). The total tax per unit levied on imported spirits increased steadily from T$55.38 per Lal in FY2014–5 to T$87.74 per Lal in FY2017–8. This resulted in an increase in tax revenue, from T$2,536,700 to T$4,204,700. Nevertheless, the amount of imported volumes decreased from 50,250 Lal in FY2016–7 to 47,920 Lal in FY2017–8, reflecting the tax increase (see Figure 72). Sources: Customs duty, excise tax, consumption tax, and total tax: sourced from the Ministry of Revenue & Customs, Tonga. 81 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 70: Tax per unit of locally manufactured spirits Figure 71: Tax per unit of imported spirits Sources: Excise tax, consumption tax, and total tax: sourced Sources: Customs duty, excise tax, consumption tax, and total from the Ministry of Revenue & Customs, Tonga. tax: sourced from the Ministry of Revenue & Customs, Tonga. Figure 72: Consumption, tax revenue and Figure 73: Consumption, tax revenue and total tax per unit of imported spirits total tax per unit of locally manufactured spirit Sources: Imported volume, tax revenue, and total tax: sourced Sources: Producing volume, tax revenue, and total tax: from the Ministry of Revenue & Customs, Tonga. sourced from the Ministry of Revenue & Customs, Tonga. The total tax per unit levied on locally The total tax per unit levied on spirits manufactured spirits increased steadily increased steadily from T$52.54 per from T$24.15 per Lal in FY2014–5 Lal in FY2014–5 to T$79.62 per Lal in to T$34.50 per Lal in FY2017–8. This FY2017–8. This resulted in an increase resulted in an increase in tax revenue, in tax revenue, from T$2.64 million to from T$110,700 to T$297,500. The T$4.50 million. However, the sales volume production volume has also increased decreased from 58,060 Lal to 56,540 Lal during the same period (see Figure 73). (see Figure 74). This could be due to the fact that locally produced spirits were subject to lower proportion of tax increases, and hence they remained cheaper than imported spirits. 82 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga There was a reduction in the frequency Figure 74: Consumption, tax revenue and total tax per unit of spirits with which spirits were consumed. Prior to the FY2017–8 tax increase there were more drinkers consuming spirits three to four times and five to six times a week. Following the tax increase, the number of drinkers in these frequency groups fell to just 1.42 percent and 0.41 percent respectively. There were more drinkers shifting behaviors to drink once a month or rarely (see Figure 75). Effects of consumption tax exemption on imported Sources: Sales volume, tax revenue, and total tax: sourced healthy products from the Ministry of Revenue & Customs, Tonga. Figure 75: Consumption frequency of spirits The Government of Tonga introduced consumption tax exemption on a number of products perceived as healthy in July 2016 in anticipation that the price of these products would fall and their consumption would rise. Key products subject to consumption tax exemption in July 2016 are listed in Figure 76. Nevertheless, very few people were aware of the consumption tax exemption policy. Only 12.6 percent of consumers were aware that the government had exempted imported fruits and vegetables from consumption tax and even fewer Source: Baseline and endline survey data (2017). were aware of the tax exemption on other food products. It is also interesting to note Figure 76: Public knowledge of food products subject to tax exemption, July 2016 that not all products introduced for the tax exemption in July 2016 were healthy, e.g. sausages and ham. The World Bank research team decided to focus on imported fruits, which are easily accessible to Tongans from markets, and used the household survey as the main tool to assess the impact of consumption tax exemption policy. Apples and oranges are the most common and accessible imported fruits in Tonga, and were selected as the proxy of imported fruits. Source: Baseline survey data (2017). 83 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Apples Consumption tax exemption was not passed on to consumers – prices stayed the same despite consumption tax exemption. Before the tax exemption, a bag of four or five apples apples cost T$5 per bag. Following the tax exemption, 94 percent of Tongans reported that the price remained the same (see Figure 77), while nearly half stated that the amount of apples decreased (see Figure 78). Picture 8: Apples sold above maximum price set by the Competent Authorities Figure 77: Change in price after the tax exemption policy Figure 78: Quantity of apples following the tax exemption policy Source: Baseline survey data (2017). Source: Baseline survey data (2017). Most people did not change behaviors Figure 79: Apple consumption behaviors in consuming imported apples. As the price did not change, 75 percent of Tongans’ apple consumption behaviors also remained the same, and approximately 23 percent decreased their apple consumption. Several shops sold apples above the maximum price set by the Competent Authorities. The maximum price for apples set by the Competent Authorities for July 2018 was T$3.89 per kg. However, the price of apples sold in markets was found to be above this maximum price during the same period. As an example, a shop in Nuku’alofa sold the bag of apples in Picture 8 for T$5, though the Source: Baseline survey data (2017). bag weighed only 950 grams. 84 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Oranges Most Tongans did not change their behaviors in consuming oranges. Like apples, oranges in Tonga are sold by the bag and not weighed as shown in Picture 9. While the number of oranges inside a bag may have varied, the price of a bag of oranges did not change. It was T$5 before the tax exemption policy, and remained the same after the tax exemption policy. Several shops sold oranges above the Picture 9: Oranges as commonly sold in Tonga – maximum price set by the Competent in bags and not weighed Authorities. The maximum price for oranges set by the Competent Authorities for July 2018 was T$5.7 per kg. However, the price of oranges sold in markets was found to be above this maximum price during the same period. As an example, a shop in Nuku’alofa sold the bag of oranges in Picture 10 for T$5, though the bag weighed only 600 grams. Picture 10: Oranges sold above the maximum price set by the Competent Authorities 85 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Views from Communities 86 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga This chapter presents key findings came into force. Those who learned from focus group discussions with about the tax policy from the radio said communities in Tongatapu and Vava’u, they felt that the broadcast information did and takes into account gender, not make it clear which goods were taxed urban/rural, and socio-economic status and which were not. It was suggested that aspects. the information about NCD tax should be communicated more clearly and not only Food through television and radio, but also via Facebook, the churches, and newspapers. The majority of rural and urban women Facebook was seen as an important form and men in Tongatapu and Vava’u of communication. Reference was made were not aware of the NCD food tax to a Facebook page called “Tonganow”. implementation in 2016 and 2017, Other suggestions included: voicemails, or why it happened. Most said they door-to-door messaging, and village meetings. only noticed a price increase when they Village workshops with face-to-face went shopping but did not know about discussions were also proposed, in particular the new taxes. When asked what was to avoid potential misunderstandings. driving the tax reform they thought it Reference was made to the Tongan was to finance civil servants’ salaries, to Communications Corporation which uses enhance revenues to fund government mobile phones for public information alerts activities, and/or to reduce consumption such as cyclone warnings. of unhealthy foods, tobacco and alcohol. Reactions on the food tax were There was confusion and mixed. While many participants partly misunderstanding about price supported the government’s move, changes made by individual retailers, viewpoints differed between locations, with many consumers initially blaming socio-economic status, and by the retailers for raising their prices. It was availability of heathier food options shared during a focus group discussion (e.g. seafood). Where households catch in Neiafu that “Chinese shop owners fish and seafood themselves, or have decide by themselves on price increases access to affordable fish and seafood, and when they come into effect. People people welcomed the tax. While many sometimes blame the government, but see the potential health benefits and in some ways it is the Chinese business reduced mortalities (or premature people. Government is straightforward.” death) as a result of increasing tax, Similarly, an interviewed Principal many believed that an economic Enforcement Officer at the Ministry of approach alone cannot solve behavioral Commerce reported more incidents of problems related to food consumption. retailers charging more than the price rise recommended by the Price and Wage Control Act. Most of those who knew of the NCD tax policy learnt about it through radio broadcasts such as live Parliament sessions, although the communities would have liked the policy to be communicated through multiple media channels and be given an opportunity to provide feedback before the tax 87 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Box 2: The impact of food tax on consumption among different socio-economic groups Communities made it clear that the food tax increase affected poor communities in three ways. First, not so well-off households struggled to make ends meet due to the increase, which in some cases led to delayed payments of household bills. Second, the increase in the price of chicken meat and the absence of affordable fresh fish, seafood, or plant protein pushed people to turn to unhealthy alternatives (e.g. corned beef and salted beef). Third, tight household budgets did not allow households to explore new ways of eating (e.g. buying more fruit). It was reported that rising food prices were particularly difficult for families with large numbers of children. Coping mechanisms were in place for those able to prepare traditional foods or vegetarian meals and with access to home grown produce or seafoods. In comparison, well-off households could afford to continue to eat as before. Regardless of socio-economic status, all Tongan people are united by their culinary traditions which include taste preferences for fatty foods, large portions and feasts. Fruit and vegetable consumption was found to be low throughout all socio-economic groups before the tax policies. More effective nutrition education on adequate portion sizes and healthier ways of cooking (e.g. with more vegetables and smaller amounts of meat) may be the way forward to bring about positive, affordable dietary change across all socio-economic groups. Cases of positive behavior change and innovative thinking have been observed across all socio-economic groups. However, healthier dietary change usually takes place first among well-off groups (trend setters) with spill-over effects to other groups at a later stage. Many focus group discussion made from flour, sugar and oil) as coping participants in Tongatapu and Vava’u strategies. “In this way it keeps us out of were concerned with the effects of tax hardship and deficiency,” one woman said. increases on poor communities and the unemployed who cannot afford to Mutton flaps and turkey tails were eaten buy healthy food alternatives. They less often and in smaller amounts, with compared the situation with policy increased prices being the main factor measures in other countries, stating that for reduced consumption. Many stated in other countries expensive meats can be that before the price increase they were substituted with quality food. In comparison, able to buy 1–3kg mutton flaps several they claimed, Tongans do not have healthy days per week, but now they could only affordable food choices. They gave an afford to buy mutton flaps to consume on example that ex-patriates and government Sundays. Rural and less well-off men in workers can eat T$50 dishes of quality Tongatapu reported lower consumption food. “Poor and less well-off families, in of mutton only because of the scarcity turn, could only afford T$5 [poor quality] of money. “Once we have money, it is take-aways.” Not well-off women in Neiafu better for us to eat a tiny amount of mutton, reported that it is difficult to make ends rather than a kilogram of chicken or other meet, which includes late payment of bills meat.” Similar to mutton flaps, turkey tails and late repayment of loans. It was found are still popular and people would buy the situation was particularly difficult turkey tails if they could afford them: “If for families with large numbers of we got some more [money], there would children. People think of fishing and be more consumption [of turkey tails].” planting pele (a type of spinach) or It is important to note that youth eat economizing (e.g. preparing simple pancakes differently; they rarely eat mutton flaps and turkey tails. 88 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Mutton flaps and turkey tails are mainly imported noodles alike. Men in Ma’ufanga replaced by chicken, while beef and fish pointed out local noodles are not as are also alternatives. People reported popular as imported ones. that instead of buying 1 kg of mutton they would buy 3 kg of chicken, which can Sodas remain popular, yet as the price feed a whole family for the same price. increased, consumers chose cheaper The price of chicken varies along the brands. Sprite and Coca Cola (at T$2.50 retail chain. Price ranges were reported of per bottle) do not sell well anymore, but between T$2-3 in Tongatapu and T$4-5 locally manufactured SSB (e.g. South in Vava’u. Canned fish was found to be Star), which is T$1.50 per bottle, has another popular alternative to fatty meat. become very popular. In particular, children It is important to note that low-income have an appetite for flavored drinks high earners reported that they were negatively in refined sugar, and SSBs are still very affected by increasing chicken prices, yet popular during meetings and church chicken remains the cheapest meat on gatherings. the market. The majority of participants reported Instant noodle consumption has not that fresh fish has become expensive, decreased and instant noodles remain though consumption behaviors varied. popular. Women in Tongatapu said the All participants want fresh fish to price of instant noodles was still low, with become cheaper. The well-off group in one trader reporting “[they] turn over fast in Tongatapu said: “Fish is expensive, but my shop”. Yet prices are rising. Rural men we know it is safe and good. We don’t in Tongatapu reported that noodles are go for cheap, we only go for what is safe very easy to prepare which is attractive, (for consumption).” The less well-off in while not so well-off women from Neiafu Tongatapu find fresh fish too expensive reported that noodles “cause health and eat canned fish instead. If fish were problems”. Some people had stopped cheaper, the common view was that it eating them, and some said the government would be more popular. Rural women in should prohibit noodles all together. Tongatapu were not aware of the Ministry Women also suggested a return to diet of of Fisheries program to sell fish at T$5 per the past and a stop to making expensive kg. They thought the program “only benefits food choices. Nuku’alofa residents who have easy access to this program”, stating that they Children in particular are hooked on themselves had to buy fish at a higher price. noodles and artificial flavors. Previously, noodles were not common, but it Middlemen may contribute to higher was observed that as a result of the prices among healthy products. For presence of Chinese traders, Chinese example, some farmers prefer to retain foods were introduced. Over time, their crops and then export them or bring noodles became very popular, particularly them to middlemen. Not so well-off men in among children. Children eat them raw, Ha’asini said some farmers “will only sell for breakfast (as a meal), or as snack their produce locally if there is a pressing or for lunch. It was reported that children need for cash. Farmers who sell [produce] do not favor traditional foods. Focus at the market, say a bag of yams for T$20, group discussion participants gave an are bought out by middle-men (called ‘kau example: “If we offer la’ipele (Tongan hali’). They raise the prices ridiculously, up spinach) children do not like it.” By contrast, to T$50―70. These middle men should when serving imported or local noodles, be paying lots of taxes.” It was reported “they devour [them]”, both local and that introducing business licenses may 89 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga not be an easy solution as financial benefits for middlemen are high, e.g. a fine may cost just T$5 while financial gains from sales may be as high as T$100. Dietary choices are driven by taste rather than nutrient content. Well-off people in Ma’ufanga highlight that people “don’t think about nutrients. People go for sugary and salty taste”. Yet, this group in Ma’ufanga was able to give an example of a healthy and sweet food which they said is tasty: banana and oats in the morning, providing evidence that innovation in diets can occur. Having a big meal is part of Tongan culture and being obese is accepted. One respondent coined it like “eat till you can’t fit through the door”. There is also the culture to encourage others to eat “influencers” and people tend to follow more. They noticed that people in other their examples. However, traditions are countries eat differently and less heavily. not easy to change for everybody. When In Tonga, they claimed, each meal is asked how it would be to show up at a exceptionally large, and people have funeral with vegetables, it was reported an “appetite” for fatty food, which that one would be subject to ridicule: “requires fatty food to quench”. People also “Your descendants would carry the shame, reported that they grew up with mutton [and people] would throw the veggies at flaps and developed a taste for it. Overall, you.” people reported that while buying and consumption patterns may change as a While consumption of unhealthy result of price increases, the “craving” food is prevalent in Tonga, there are remained. Nevertheless, communities examples of healthy eating among also reported the generational difference some Tongans. Many have older relatives in meat preference – while adults prefer and observed that their traditional diets red meat, “children love chicken”. kept them healthy. “We didn’t hear of diabetes amongst our mothers and The church can provide a positive fathers, they lived into their 90s on meals influence on congregations. Traditionally, of crops and la’ipele (Tongan spinach) church gatherings that could last up to a boiled with coconut cream. I still have that week offered excessive food and provided meal, but my children have no appetite up to five or six meals a day. Today the for it. I tell them to eat as I do. I don’t eat practice is perceived as ‘unhealthy’ and noodles, I never will.” Another woman in recent years behavior has changed. recalled her mother eating vegetables For example, traditional feasts now serve such as plantain and sweet potato cooked food in individual bowls (instead of at with the skin. “So I have [lived] on that. I buffets), and there are vegetarian feasts, ask my children to do so. My grandmother and watermelon juice instead of SSBs. who lived to her 90s had the same meal. Church leaders and nobles are seen as The challenge is that we cannot make a 90 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga leaves with coconut cream without chicken. Let’s return to the traditional foods, what we were used to, if we can do it for our children, then diabetes will be on the decline and medications will not be needed.” Some people expressed the need to introduce culinary innovations and voiced demand for fresh local produce. A group of rural men said: “If we continue to cling to the old ways we will miss out on many things. It is important for us to examine how change is happening. We can try something new, say replace meats with taro leaves and other things.” In order to replace SSBs, some people were using natural ingredients such as kola limes, water melon, coconut, and lemongrass to make drinks. Sweetened ice blocks, coffee and tea were reported return to that kind of meal. We have now as substitutes for ice cream. Also, it was moved to an age where we are driven reported that some women had started by taste. [People] want a big piece of using eggplant, which was not considered fatty meat. If we [were] to return to the edible in the past. On the supply side, past and eat vegetables, Tonga [would] one retailer reported selling home-made be a healthy nation again.” Some also caramel sauce and coconut cream instead observed that foreigners eat a lot of of manufactured versions. vegetables. One woman reported that she started to grow vegetables for sale, People recalled the late king HM Tupou and also for home consumption which – IV and his efforts to lose weight. It is according to her – is much better than known that he undertook physical buying “all sorts of high fat meats”. Another exercise, in particular cycling. He woman said: “I challenge you ladies to died of natural causes at the aged 86 make a return to traditional foods – pawpaw years. People see this as evidence that (papaya) and coconut in the morning, taro NCDs can be addressed successfully. Box 3: Response to the food tax – urban and rural differences Urban areas The overall impact of the policy intervention was considered to be minimal for well-off urban families, while not well-off urban families lowered their consumption of unhealthy food (particularly mutton flaps, etc.). Demand for instant noodles remained the same because of the absence of a good replacement for young people. Rural areas In rural areas, the impact of the policy intervention was observed in both well-off and not well-off families. Healthy alternative food is more available to rural families compared to urban areas (families use more local foods such as taro leaves, local spinach ‘Pele’ and even seafood). 91 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Tobacco The majority of smokers across Tonga People reported unfair business reported shifting to use Tapaka practices by traders. They recalled one Tonga and buying fewer manufactured instance in 2018 when a shortage of cigarettes. This effect was more cigarette rolling papers was perceived pronounced among men and in rural areas. as a move to encourage people to buy Not many female smokers in Vava’u imported Palataisi cigarettes. Also, it reported the shift to Tapaka Tonga, as was reported that some Chinese traders they did not like the taste; instead, they placed a high mark-up on the price of bought fewer cigarettes. More women in cigarette rolling paper. It was reported Tongatapu appeared to shift to Tapaka that several Chinese-owned shops Tonga, like men. do not provide receipts and have no cash register machines, while others A pack of Tapaka Tonga lasts much do not keep sales records. Chinese shops longer than a pack of cigarettes. are reported “to set mark-ups as they Smokers in Tongatapu reported that in the please. Sugar is allowed a 10 percent past they bought two packs of cigarettes [price increase], but [they] apply 15 a day (T$14―15 per pack) but the tax percent. This contributes to price prompted a shift to Tapaka Tonga – a bag fluctuations and confuses consumers on the of which costs T$5 per pack and lasts real reasons why prices are increasing”. for one week. In Vava’u, people reported that a packet of Tapaka Tonga lasted for three days, while a pack of manufactured cigarettes would only last for one evening at a kava party. 92 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga There was mixed feedback on women’s There is a recommendation for anti- experience of the tobacco tax. Some smoking efforts to focus on youths: said they smoked when they had enough “We ‘old dogs’ can’t learn new tricks”, money, and if there were a shortage of said a man from Ma’ufanga, Tongatapu. money, they would sacrifice their The group suggested that the cigarettes in favor of buying food. Government used the resources to However, one woman reported, “my target youth, and to get youth to address children know that smoking is very the problem of teen/young adults important for me, they would rather give smoking. back to me their lunch money so I can smoke”. Another said: “I don’t care if Alcohol they don’t have lunch, I would rather smoke and stay at home and get my work Overall, consumption of beer and spirits done, if not I won’t get anything done as I has fallen. Occasional drinkers and really need my cigarettes.” daily drinkers reacted differently to the alcohol tax. Occasional drinkers reduced It was also reported that smoking caused their consumption while daily drinkers marital problems. For example a woman hardly changed their consumption from Tongatapu said: “Because my behaviors. husband can sacrifice in order for our family to have food, I will return from Alcohol consumption needs to be seen the shop with a pack of Pall Mall. And in the larger social context and requires then we have an argument. They can go more research. Though the overall ahead and eat, I will sit and smoke, that’s purchase of alcohol has fallen, some more important.” Similarly, another woman people who do not (or cannot) buy stated that “the cigarette is like a big alcohol themselves were offered or given and important thing for me. If we were to drinks by other people. For example, record our daily shopping, we would those on low incomes who want a drink notice that cigarettes make up a big portion rely on friends to provide it. This practice of the expenses even to the extent of is common and has the name “piggybacking” cutting down on family needs in order to (‘namu’). Also, some said alcohol save money for a pack of smokes”. consumption is not on the decline, but rather increasing and needs to be better Many smokers did not appreciate surveyed. The increased consumption anti-smoking TV ads. Some women is not observed in people drinking in believed that everyone already knows groups (or public), but rather in private, the risks to health, but they still smoked: e.g. in their vehicles or at home, as “It is a waste of time and effort to have reported by men from Ha’asini. programs to create awareness.” The Information on youth- and alcohol- concept of ‘let’s eat and let’s talk’ is related offences could also be sought. preferred. “Now”, one woman explains Participants also questioned why “when there’s something on TV, I turn the government functions still plentifully TV off and go do something else. That’s served alcoholic drinks. the truth.” Another woman said: “I could damage the TV for broadcasting anti- tobacco programs. My father has the same attitude.” 93 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga For men, kava is a main substitute. or “say offensive words to parents”). Urban, well-off men said that in the past, Increased tobacco prices have also seen alcohol was cheap and affordable, while younger people shifting to sniffing glue these days, “one is more selfish. Say for bought at repair shops. Glue sniffing me, I will buy just one drink and leave it at was reported also among primary students. home, then I’ll go to the kava party and drink Meanwhile, other Tongans have switched kava. At the end of that session, I’ll go home to smoking Tapaka Tonga. and finish off with my beer”. And though the price of kava Tonga has risen from T$25 to T$150 per kg, consumption has not fallen. Future of NCD tax There are also reports of the uptake of Most participants in the focus group dangerous substitutes. discussions did not support further Some people have switched to home increases in food taxation, mainly for brewed alcohol, while others have economic reasons. The few who did switched to marijuana and ice (crystal express support mentioned health as the methamphetamine). Under the influence main reason. Non-smokers are generally of ice, people reported they do not really supportive of the tobacco tax, while smokers know what they are doing (e.g. they are not satisfied with the increased tax on could “commit murder without knowing” tobacco. Box 4: Impact of taxes on retailers Most retail stores had heard about the tax increases in July 2016 and July 2017. They heard about the tax via newspapers and television. All retail stores increased their price mark ups as a result of the tax. However, they also reported difficulties in maintaining their customer base while having to raise prices. Several retailers shared that they faced challenges of fluctuating prices, which were passed on to them by Chinese middlemen. Most retailers faced reduced sales of key products that were subject to the tax increase, particularly manufactured cigarettes, mutton flaps, turkey tails, and ice cream. According to small retail stores, they used to sell one carton of cigarettes per day prior to the tax increase, but “nowadays one carton lasts two days.” Another retail store reported that before the tax they sold one carton of mutton flaps a day, “now it [one carton of mutton flaps] lasts two days”. At the same time, sales of Tapaka Tonga have risen significantly. Retailers also reported that customers brought significantly fewer mutton flaps and turkey tails, so they had to stock fewer of these products, and that customers substituted mutton flaps and turkey tails with chicken. Several retail stores reported that the NCD tax negatively impacted their sales and incomes. A retail store in Tongatapu said: “[Tax] had a big impact. When prices increase, customers leave and goods don’t turn over, and we don’t get profit. Our cash flow is slow.” Several retailers in Tongatapu reported that customers thought they (the retail stores) raised the price for their own ends, and “asked why”. While sales of many products that are subject to NCD tax dropped, sales of cheaper goods that are not subject to NCD tax were unaffected. While NCD tax made cigarettes, mutton flaps, and turkey tails more expensive and their sales dropped, sales of cheaper products such as imported instant noodles remained high, as their price is still low. 94 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga While the NCD tax is generally not • Support for domestic vegetable production popular, the communities would be or chicken farms. more supportive if the government • Support for medical equipment, improving could demonstrate that revenue from medical service delivery. the NCD tax was going to the following • Developing public sports facilities and areas: annual health weeks/programs. • Support for health-promotion activities • Support for retirement benefits. and making healthy food cheaper, including subsidizing the price of fish. Communities advised that the following • Engagement with the private sector to can have a strong (either positive make healthy food more accessible and or negative) influence on healthy providing tailored support to Tongan lifestyles: retailers to help them stay competitive. • Cultural obligations and traditions • Creating scholarships for young people • Church leaders to have advanced study abroad and come • Community leaders and nobles back to serve the country. Box 5: Women’s and men’s perspectives on NCD tax Food Female focus group discussion participants are a lot more eager than men to introduce healthy diets to their families if healthy choices are made available through government support. Male participants are more curious to know how the NCD tax revenues are used by the government, and in what way that will allow for more healthy diets and choices for the public (e.g. subsidizing the fuel of local fishermen). In rural areas, female participants are very specific in their needs for family empowerment to introduce healthy behaviors such as providing cooking classes, supporting home gardening, chicken and sheep farming. Male participants would like to see more visible government commitments in terms of supporting rural communities, specifically on health education, benefits from the Ministry of Fisheries’ healthy initiatives, support for subsistence farming and greater engagement in discussion of health policy interventions. Tobacco and alcohol consumption Policy interventions yielded limited benefits for well-off women in both urban and rural areas. Higher priced tobacco and alcohol products are not a significant barrier for well-off women regardless of where they live. These products are still affordable for well- off women, while the Tobacco Quitline support services, in their current form, may not be appealing to well-off women. There are differences in the shift to Tapaka Tonga between men and women. Significant numbers of men, regardless of their socio-economic status and where they are from (both Tongatapu and Vava’u) shifted to Tapaka Tonga, though this affected less well-off men more. In Vava’u, more men shifted from manufactured cigarettes to Tapaka Tonga. Female smokers in Vava’u appeared to resist the shift to Tapaka Tonga as they did not like its strong taste. However, in Tongatapu, more women also shifted to Tapaka Tonga, though well-off women kept smoking manufactured cigarettes. Men are more inclined to look for and use alternatives for alcohol products such as homebrew, sniffing glue (outer island boys at high-school age), and drugs (youth in both rural and urban areas of Tongatapu). Some women and men are prepared to sacrifice their family’s financial commitments such as school fees, bills etc. in order to fulfill their personal desire for tobacco and alcohol. 95 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Discussion 96 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Positive impacts from the NCD tax Tapaka Tonga is the key unhealthy policy have been observed. The excise substitute to manufactured cigarettes; tax resulted in price increases in all taxed and is cheap and widely accessible products, made cigarettes less affordable in shops. According to the Ministry of and affected the behaviors of smokers. Revenue and Customs, over 90 percent Tax on cigarettes had a greater effect of Tongan shops sell Tapaka Tonga. on “less well-off” smokers, as more less According to interviews, retailers pay well-off smokers reduced consumption of manufacturers just T$2-3 per pack of manufactured cigarettes. This is consistent Tapaka Tonga, and retail it to consumers with the Lancet findings on the equity impacts at a price of T$6-7. Tapaka Tonga is of price policies. Price, rather than health, currently not taxed and is very profitable is the main reason among those who to retailers. decided to change behaviors and reduced consumption. Increased excise tax has led to a decrease in the frequency and amount Nevertheless, the positive impacts of alcohol products consumed, of the tobacco tax policy in Tonga however, a shift to substitutes has have been diluted by the availability been observed. Significant numbers of cheaper substitutes. There was a of drinkers have shifted to kava. There shift from more expensive brands of is also an increased incidence of cigarettes to cheaper brands, particularly substance abuse, including marijuana, locally manufactured cigarettes that are glue/petrol sniffing, and crystal subject to less tobacco tax. But more methamphetamine, though it is not clear importantly, there was a big shift from whether this is related to the increased manufactured cigarettes to the untaxed excise tax. Tapaka Tonga. Today, most tobacco users in Tonga smoke Tapaka Tonga. Food consumers do not always respond to price increases. While Tapaka Tonga is no less harmful to excise tax on turkey tails, mutton flaps, health than manufactured cigarettes. and ice cream helped reduce consumption Many Tongans misunderstood that of these products, it had very limited Tapaka Tonga is less harmful than effects on the consumption of chicken leg manufactured cigarettes, and some quarters. In addition, the tax on chicken even thought that it was “healthy” given leg quarters appeared to have regressive that it has been marketed as “organic effects on less well-off households. tobacco”. However, according to WHO Despite the price increase due to tax, analysis, like manufactured cigarettes, chicken leg quarters remained the Tapaka Tonga contains many toxic cheapest meat in the market, and less chemicals, including nicotine, formaldehyde, well-off households did not have access acetaldehyde, acetone, acrolein, to cheaper or more healthy options in the propinaldehyde, crotonaldehyde, market. Consequently, the tax on chicken butryraldehyde, and 2-butanon. The leg quarters took a proportionately greater amount of nicotine in Tapaka Tonga is amount of financial resources from those proportionately higher than that of with lower incomes. manufactured cigarettes.33 33 Health Science Authority of Singapore report, 2 April 2013 (obtained from World Health Organization). 97 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Lack of affordable healthy food The promise of changes in diet as a alternatives remain a major issue in result of excise taxes on unhealthy Tonga. More policy interventions are products is there, but more needed to make healthy food more dramatic changes in local diets are affordable, accessible and tasty. For needed. The findings from the focus group example, fresh fish is highly desirable, discussions suggest an overall reduction but less well-off households reported in the quantity and frequency of the that they cannot consume fresh fish on a consumption of mutton flaps, turkey tails, regular basis due to the high price. Canned and a general shift toward the consumption fish, which is currently consumed instead of chicken leg quarters and tinned fish. of fresh fish, often contains added oil However, to effectively address NCDs, and salt and the lining in cans could also a more significantly reduced intake of include unhealthy chemicals (e.g. BPAs).34 calories, total fat, animal-based protein, refined sugar, and salt is needed, together The exemption of consumption tax with an increased intake in polyunsaturated on selected healthy products (e.g., fats, fiber, vitamins, and phytonutrients imported fruits) did not result in lower (see Table 15). retail prices or higher consumption of these products. The price of imported fruits remained the same, despite the removal of the 15 percent consumption tax, meaning that the benefits of the tax exemption went to traders, not consumers. 35 Table 15: Observed and desired behavior change and dietary outcomes Observed behavior change Dietary outcome Reduction in quantity and frequency of consumption Likely reduction in intake in saturated fats, however, with of mutton flaps and turkey tails and general shift toward the continuous consumption of large amounts of animal- consumption of chicken leg quarters and tinned fish based food on a daily basis, overall intake of calories, total fat, protein and salt would need further reduction Limited reduction in quantity of consumption of ice cream, Small reduction in caloric intake, in particular of sugar SSBs and some increased consumption of local juices, water and coconut water (e.g. at church meetings) No change in consumption of fresh fruits and vegetables No increase in intake of fiber, vitamins, minerals, etc. Desired behavior change Dietary outcome Significantly reduced consumption of animal-based protein, Significant reduction in intake of calories, saturated removal of fatty animal parts (e.g. skin), and increased fat, animal-based protein and increased intake of consumption of fresh fish or plant-based protein polyunsaturated fatty acids alternatives (e.g. beans) – paired with healthy vegetable oils and minimal salt used during food preparation Increased consumption of fresh fruits and vegetables, Increased intake of fiber, vitamins, minerals, and in particular of green leafy vegetables – all year round phytonutrients (e.g. antioxidants) Significant reduction in consumption of manufactured ice Significant reduction in caloric intake, in particular of cream and SSBs and increased consumption of locally sugar and trans-fatty acids made sweets and snacks with local low caloric ingredients (e.g. ice cream made from fruits) as well as of local fruit juices or herbal infusions 34 Bisphenol A (BPA) has been shown to affect lipid metabolism and promote weight gain in animal studies, but also human populations. See for example, Health Promot Chronic Dis Prev Can. 2017 Dec; 37(12): 403–412. 35 The makers of this policy wanted imported fruits to be affordable so that school children can buy an apple for T$0.30-0.40 instead of instant noodles. Tonga has a price control mechanism for imported fruits; however, it seems that the policy has not been effectively enforced. 98 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The design of the excise tax policy the fact that the list of products to be in 2016 was led by the Ministry of subject to the NCD tax policy in July Revenue and Customs, with limited 2016 was suboptimal. Some products involvement of other relevant which were not supposed to be “unhealthy”, stakeholders. According to in-depth e.g. chicken, were included in the list of interviews with policymakers, the Ministry products to be subject to tax, while many of Revenue and Customs took the lead unhealthy products with much higher in designing the new excise tax policy levels of salt, fat, and sugar were not introduced in July 2016, including included. At the same time, products identifying which products to tax. There not considered “healthy”, such as ham were limited consultations with experts and sausages, were included in the list and key stakeholders within the government of products to be tax exempt. Furthermore, and outside during the tax design phase. MCCTIL was not adequately consulted Several high-level officials from relevant during the tax design phase, though the ministries (e.g. MOH, MAFF, MCCTIL) government expected the ministry to help reported that they were not consulted monitor the price of products that were during the design phase, and only learned subject to exemption. about the proposed tax policy when it was submitted to the cabinet for review and Throughout this study the World approval. Bank research team has highlighted the importance of consultations with Lack of proper consultation with key relevant experts and key stakeholders stakeholders and experts, particularly during the design phase. This not only nutrition experts, during the design supports evidence-based tax policy design, phase may have limited the positive but also supports the implementation health impacts that the policy could of the tax policy, e.g. in price monitoring. have generated and misleadingly sent As a result, consultations with key ministries the message that the government was and stakeholders during the tax design more interested in revenue generation phase in the later years has been than health. This was demonstrated by significantly improved. 99 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga How much did consumers respond to tax-induced price changes? Consumers’ responsiveness to changes Figure 80: Elasticity of taxed products in price led by the imposition of excise tax/import duty varied from product to product. The World Bank research team conducted a quantitative measurement of consumers’ responsiveness to changes in price led by the imposition of tax. The measure is based on the consumption elasticity concept; it is calculated as the percentage change of consumption relative to the percentage change in price. The pre- and post-tax data were taken from before July 2015 to June 2016 and July 2016 to June 2017, respectively. Import volume data were based on the import volumes provided by the Ministry of Revenue and Customs. Price data were Source: Endline survey data 2017. drawn from the consumer price index and retail surveys. Figure 80 reports elasticity of the taxed products. As can be seen in Figure 80, tax- Smokers from lower wealth quintiles induced price changes caused are more responsive to price changes Tongan consumers to reduce their and are likely to reduce cigarette consumption of taxed products to consumption as the price of cigarettes differing degrees. Instant noodles and increases. Figure 81 provides elasticity chicken legs had price elasticity of -0.38 values of cigarettes by wealth quintile. and -0.39 respectively. The negative sign The estimates are based on baseline and indicates that price and consumption vary endline survey smokers reporting inversely. Thus, for instant noodles, one “decreased” and “no change” in their percent increase in price will reduce instant smoking. Prior to the calculation of noodles consumption by 0.38 percent; elasticities, the value of 0 elasticity was similarly, 1 percent increase in the price assigned to those whose smoking did not of chicken legs will reduce consumption of change. Quintile-wise, elasticities showed chicken legs by 0.39 percent, which shows that a more affluent smoker was less that consumption of these products did not responsive to price change. In addition, change significantly compared to changes elasticity figures from endline data were in price. Elasticity of SSB is -0.93 which is more responsive to price changes. This within the ranges of WHO’s findings and is implied that as time went by, consumers more sensitive to price change compared began to reduce their smoking frequency to chicken legs and instant noodles. Two – a result consistent with the aims of this products most responsive to price change tax initiative. are ice cream and mutton flaps, with elasticity of -1.38 and -1.92 respectively. 100 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 81: Elasticity of cigarettes Source: Baseline and endline survey data 2017. While it cannot be concluded that these products, therefore, would be less the tax increase was the sole reason sensitive to price change. Products whose for the reduction in consumption of unit price occupied a significant proportion taxed products, the tax can be said of household expenditure would be more to have had significant influence on sensitive to price change. Elasticity also the consumption pattern of Tongan depended on the nature of the product consumers. For example, around 76 itself. Products considered as a daily to 79 percent of smokers said that they staple would be less sensitive to price lowered their smoking because of the change. Luxury products would be more price change (see Figure 15). Similarly, sensitive to price change. Lastly, addictive around 98 to 99 percent of mutton flap products such as cigarettes would be less consumers reduced their consumption sensitive to change in price. When wealth because of the change in price (see is taken into account, smokers from lower Figure 33). The fact that, during the period wealth quintiles are more likely to reduce of this study, there was no other variable their consumption behaviors as price that caused significant change prices on increases. The results shown in this these products also lends support to the section provide empirical support for claim that the tax was the major driver for the effectiveness of the use of tax as an lower consumption of the taxed products. instrument to divert Tongan consumers away from consuming products that are To conclude, how much a consumer detrimental to their long-term health and responded to price change depended well-being. on the characteristics of the product. For a product whose price per unit was low, a change in price would not have significant impact on total consumer (and household) expenditure. Consumption of 101 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga What are the key substitutes to taxed products? The potential risks of swapping to other, Based on the endline household survey cheaper and/or low-quality unhealthy data, while consumption of mutton food products are substantial, given flaps reduced significantly due to the variety of food choices in Tonga. higher prices resulting from the tax While unhealthy food products that are policy, over 40 percent of respondents subject to excise tax have become more reported shifting to alternative, cheaper expensive, there are still unhealthy food products. The most popular alternatives products that are not subject to excise were salted beef, followed by tinned tax, making them more affordable than the fish and corned beef (see Table 16). It taxed products. Therefore, it is possible is important to note that salted beef and that people may switch from the more corned beef were not subject to excise expensive taxed products to cheaper tax at the time of this survey, but they unhealthy products not subject to NCD tax, also contained high and unhealthy levels diluting the effects on health outcomes, of salt and fat. For turkey tails, consumers which are the primary goal of NCD tax chose more healthy products like tinned policy. Hence, it is crucial to monitor the fish and chicken. The excise tax on SSB substitution effects of the taxed products. also led 23 percent of respondents to go for an alternative healthy product – water, the main substitute of SSB. Table 16: Key substitutes of taxed food products Taxed products Key substitutes Mutton flaps Salted beef, tinned fish, corned beef Chicken leg quarters Tinned fish Turkey tails Tinned fish, chicken SSB Water Ice cream Local ice cream Instant noodles Local instant noodles Source: Endline survey data (2017). 102 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga It is important to note the main the prices of illicit products are lower substitutes for ice cream and instant than the retail price of legal products, noodles were identified by respondents in order to make them more attractive as locally manufactured ice cream to consumers.36 Hence, illicit trade is and locally manufactured instant one of the major risks that could noodles. Before the imposition of excise undermine the purpose of the NCD tax tax on ice cream and instant noodles, to reduce consumption of unhealthy nearly all ice cream and all instant noodles products for better health outcomes. in Tonga were imported, and hence the tax was only imposed on imported ice Given the significant excise tax being cream and instant noodles. This provided imposed on a number of unhealthy a large loophole that allowed local products in Tonga, the risks of illicit businesses to set up small factories trade, particularly smuggling of tobacco manufacturing instant noodles and ice products, is substantial. According to cream. As these locally manufactured interviews with Tevita Lavemaau, former products were not subject to tax, it became Minister for Revenue and Customs and cheaper than the taxed, imported ‘Anisi Bloomfield, former CEO of Ministry of products. This demonstrates the importance Revenue and Customs during the qualitative of imposing unified tax rates for both study, smuggling of cigarettes was high in imported and locally manufactured the past, but not much in evidence in recent unhealthy products. Otherwise, people years. According to ‘Anisi, it changed can easily shift to consuming locally because of the efficiency of Ministry of manufactured unhealthy products, Revenue and Customs staff, and high defeating the whole purpose of improving penalties. health outcomes’ with ‘the tax strategy. Nevertheless, there are still potential Illicit trade risks of smuggled goods and black markets, including from offshore Illicit trade refers to any practice related fishing boats, container ships and to distributing, selling, or buying yachts that visit Tonga for brief periods. products subject to excise tax that is The Ministry of Revenue and Customs prohibited by law, including tax has a large scanning machine to scan evasion (sale of the products without containers at the airport and port of entry. payment of applicable taxes), According to business leaders in Tonga, counterfeiting, disguising the origin there might still be smuggling of cigarettes of products, and smuggling. Illicit trade into Tonga. However, the amount tends to can be undertaken both by illicit players be small, and to be Chinese-made who are not registered with relevant cigarette brands Zhonghua and Double government agencies, as well as by Happiness, which are not popular among legitimate entities whose business Tongans but are popular among the small operations are contrary to applicable Chinese community in Tonga (around laws and regulations. In most cases, 3,000 Chinese people live in Tonga). 36 World Bank. 2019. Confronting Illicit Tobacco Trade : a Global Review of Country Experiences (English). WBG Global Tobacco Control Program. Washington, D.C. : World Bank Group. 103 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga How NCD tax revenue is used One of the most important concerns tobacco tax revenues were allocated to about the NCD tax policy raised by health, particularly to fund universal health communities is how the revenues are care, upgrade medical facilities, and train used. In Tonga, all revenues generated by doctors and nurses. Within the first year the NCD tax policy go to the government’s of introducing sin taxes in the Philippines, general revenue, and it is up to the the country raised more than US$1.2 Ministry of Finance and National Planning billion and allowed the Philippines to to decide how funds are used. provide health care to an additional 14 million families, or roughly 45 million Filipinos. Communities across Tonga have made Nevertheless, there have also been it clear that they want the NCD tax policy challenges. The Government’s capacity revenues to return to the communities, cannot absorb the sudden increase in particularly to support and promote the budget for universal health care, and health. The NCD tax policy, particularly hence structural reforms in the health the food side, is not popular among service delivery systems is required communities as it made food more before the increased coverage will lead to expensive. However, communities expressed better access to quality health services.37 that they would be convinced about the usefulness of the tax if the government While revenue from the NCD tax has could demonstrate that the revenues are increased significantly in FY2016–7 used to support and promote communities’ and FY2017–8, particularly from the health and their lifestyles. Some popular food tax, there is no evidence of a examples of how the NCD tax revenue proportionate rise in the funds is used include making healthy food allocated to health care. According to options cheaper, strengthening health Figure 82, the proportion of total Ministry promotion and disease prevention of Health budget out of total Government activities, supporting production of healthy of Tonga (GoT) budget dropped from 13 products, and improving health care percent to 11 percent in FY2014–5 and services, etc. has remained so for the past four years. Many countries have used the The proportion of the government revenues generated from sin taxes or budget that goes to health promotion NCD-related taxes to support social has always been very small, and has and health goals. For example, Thailand fallen in the past four years, despite imposed a surcharge of 2 percent above higher NCD tax revenue. Tonga Health the existing excise taxes on tobacco and Promotion Foundation is the key public alcohol to support health promotion fund. authority leading multi-sectoral health In the Philippines, revenues from the 2012 promotion and disease prevention work tobacco and alcohol excise tax reform in Tonga. Its budget allocation has always are also earmarked (soft earmark) for been low, and in recent years has actually specific programs. While 15 percent of decreased. In FY2018–9, government additional tobacco tax revenues is allocated funds for the Tonga Health Promotion toward programs to help tobacco farmers Foundation accounted for only 0.16 and workers find livelihood alternatives, percent of total government budget. 100 percent of additional alcohol tax revenues and 85 percent of additional 37 World Health Organization. Sin Tax Expands Coverage in the Philippines,’ May 2015, http://www.who.int/features/2015/ncd-philippines/en/ 104 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Figure 82: Annual budget of the Government of Tonga (GOT) and Ministry of Health Source: GOT published budget statements, Package of Essential Health Services and Sustainable Health Financing Workshop, Nukua’lofa, 29-30 August 2018. Table 17: Annual budget of the Government of Tonga (GOT) FY 2015–6 2016–7 2017–8 2018–9 Government of Tonga (GoT) 263,700,500 285,460,800 344,557,300 365,982,600 budget (T$) Tonga Health Promotion Foundation (GoT allocation) 500,000 500,000 600,000 600,000 (T$) % of Tonga health budget per 0.19 0.18 0.17 0.16 total GOT budget Source: GOT published budget statements. The increase in revenue from the NCD also be explored, including considering tax – particularly from food, SSBs, and lessons learned from different countries. alcohol tax – has increased fiscal space This is all the more important given the that will facilitate the Government of uncertainty and potential decreases in Tonga to increase the resources to external funding for health from development support health promotion/disease partners introducing earmarked taxes for prevention activities as well as improving health may also be explored, including health care services, as appropriate. considering lessons learned from different The possibility of increasing the overall countries. This is all the more important health and health promotion budget and/or given the uncertainty and potential introducing earmarked taxes for health may decreases in external funding for health from development partners. 105 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Conclusions and recommendations 106 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Conclusions The implementation of NCD-related The potential risks of substitution with taxation policy in Tonga has led to other, cheaper unhealthy food products a decline in consumption of almost are substantial, given the variety of all taxed items. This shows that most food choices in Tonga. While unhealthy Tongans are sensitive to prices. Tonga’s food products that are subject to excise tax status as an isolated economy with a have become more expensive, there are very small manufacturing sector may still unhealthy food products that are not also limit the availability of substitutes. subject to excise tax, making them more The government has also acknowledged affordable than the taxed products. Forty that smuggling of taxed goods has not percent of mutton flaps consumers reported been a major issue in recent years. shifting to substitutes – salted beef, tinned fish and corned beef. Salted beef and What drives behavior change is the “price corned beef were not subject to excise shock” following a large tax increase. tax, but they also contained high and Household surveys provide a clear unhealthy levels of salt and fat. The main indication that “price” is the most substitutes of ice cream and instant noodles important factor for changing consumption were locally manufactured ice cream and behaviors. After the large tax increase, instant noodles, which were not subject small incremental increases can be used to excise tax at the time of the survey, to sustain these impacts. Small, making them cheaper than their imported incremental increases in tax alone counterparts. This demonstrates the (without a prior large tax increase) do importance of imposing unified tax rates not work, as consumers can easily for both imported and locally manufactured adjust their behaviors to the incremental unhealthy products. Otherwise, people could price increase. The government needs to easily shift to consuming locally manufactured monitor the implementation of the NCD unhealthy products, defeating the health tax regularly and make sure that annual outcome purpose of the tax. tax increases well exceed increases in consumer prices and incomes. The increase in revenues from the NCD tax, particularly from food, SSBs, The selection criteria for food products and alcohol tax has increased fiscal upon which NCD tax was imposed in space that will help the Government of Tonga is not consistent. The use of Tonga to increase the resources to thresholds based on nutritional content support health promotion and disease could be an option to avoid this. A good prevention activities. Improving example is Chile, which applied thresholds health care services as appropriate, for sodium, fat and sugar for processed to help the country achieve universal food products.38 While the thresholds health coverage. As highlighted in the are used for labeling and advertising Pacific NCD Roadmap, such complementary purposes, they serve the objective measures are essential to address the of clearly identifying unhealthy foods. NCD crisis. This is even more important in Samoa also develops Nutrition Profiles light of the findings of this study, which show with thresholds of fat, saturated fat, salt, that without complementary measures, and sugar, in order to identify unhealthy tax increases may merely result in a switch food products. 37 Global Delivery Initiative (http://www.globaldeliveryinitiative.org/sites/default/files/case-studies/wb_food_labeling_chile_2-1-19_web.pdf). 107 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga from one form of unhealthy consumption to (e.g. by removing the fatty skin, not deep-fried). another. There are lessons learned from Chicken meat is an important source of several countries that use the revenue protein for Tongan children and adults from sin taxes to support health promotion and remains the cheapest meat on the and health services, which help these market. Most households continue to buy countries accelerate the achievement of chicken leg quarters despite the price universal health coverage. increase. The study findings suggest regressive effects from the tax on chicken Recommendations leg quarters on poorer households. Tax structure Unified excise tax should be applied to both imported and locally produced Commercialized Tapaka Tonga should unhealthy products (at the same level), be subject to excise tax as it is cheap, or its effects will be diluted. For example, widely available and harmful to health. excise tax on SSBs had earlier applied The fact that Tapaka Tonga is not subject to only imported SSBs. Local businesses to excise tax also contributes to the were quick to note the loophole and start reduction in government revenues from to produce SSBs locally. Locally produced excise tax on tobacco products, as many SSBs were not subject to NCD tax and smokers have substituted the more have become much cheaper than imported expensive, taxed manufactured cigarettes SSBs, leading to higher consumption of with untaxed Tapaka Tonga. locally produced SSBs. This also applied to instant noodles, whereby imported The government should take a consistent instant noodles were subject to excise approach and apply scientifically taxes, while locally manufactured instant supported criteria across all food groups, noodles were not, and have become including the use of nutrient profiles, in cheaper than imported ones. In addition, designing the NCD tax policy, in close the experience of imposing different excise consultation with health and nutrition tax rates on imported and locally experts from the Ministry of Health, manufactured cigarettes showed that Ministry of Agriculture, Forestry and using differential rates enlarged the Food, Ministry of Fisheries, Tonga price gap between imported and locally Health Promotion Foundation, the National manufactured cigarettes, leading more Food Council, and relevant development smokers to switch to the cheaper locally partners.This should be applied to both manufactured cigarettes. It is recommended food products subject to the increase in that excise tax rates for imported and locally excise tax, as well as food products manufactured unhealthy products should exempt from the NCD tax. Policymakers be at the same level for the same products, can ask questions such as why impose tax e.g. imported cigarettes and locally on chicken leg quarters while corned beef manufactured cigarettes. This will also and salted beef were not taxed? And why support Tonga’s commitments to the World exempt consumption tax on sausages and Trade Organization. ham, which are not considered healthy food products? Monitoring NCD tax implementation The excise tax on chicken leg quarters More effective monitoring of the price should be removed. There is a lack of of healthy food products subject to tax global evidence that meat from chicken exemption is needed to ensure that the legs is “unhealthy”, as it is more about policy of tax exemption has achieved how they are cooked to make them healthier the goal of price reduction, and in turn 108 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga higher consumption of healthy food need to have an evaluation plan before products. Most shops in Tonga do not use the policy is put in place to support the weighing scales for fruits and vegetables, collection of baseline data before making it difficult for consumers to assess implementation. if the price per unit has decreased or increased. Use of weighing scales for In addition to strengthening monitoring selling fruits and vegetables should be and evaluation mechanisms, further enforced, and it will not only help the research is needed to fill knowledge government in monitoring prices, but gaps and to inform future design and also empower customers in ensuring implementation of food taxation policy. that traders do not overcharge. The For example, currently very limited data empowerment of customers in this about Tapaka Tonga use is available. way is essential, given the fact that the More data needs to be collected to government may not have sufficient understand the number of Tapaka human resources to conduct price Tonga users, the frequency and quantity of monitoring regularly throughout the country. Tapaka Tonga consumption, and Tapaka The government should also strengthen Tonga sales volumes. This will enable measures to address overcharging across forward-looking scenario simulation all islands effectively (digitized price analysis to inform the possible improvement information systems, equipping officers and expansion of current policies. More with weighing scales, validating urban research in the areas of behavioral price lists in rural contexts, monitoring economics, nudging, as well as social road stalls, etc.). and behavior change communication may also be useful to promote consumption The Government needs to have a clear of healthy products and decrease monitoring and evaluation framework consumption of unhealthy products. for the implementation of NCD tax and More studies to understand factors and to make sure that annual tax increases determinants that can influence the exceed increases in consumer prices change of social norms to promote healthy and incomes to ensure effectiveness. lifestyles will be beneficial. Evaluation of interventions or programs needs to be planned and designed before There is room for improvement on the start of interventions. Evaluation will communication, through relevant generate evidence on impacts of the media channels, in particular mass interventions and programs as well as media and social media as well as to lessons learned. But most importantly, organize public consultations to reach the government could use available data out to different stakeholders regarding to monitor the change in prices and import/ the NCD-related taxation policy. From sales volumes of taxed products and the household survey, there was a lack potential substitutes. Existing import volume of knowledge among the majority of the and tax collection data from the Ministry of population about the imposition and Revenue and Customs could provide a proxy increase in NCD-related taxation that for sales volumes. The CPI exercise became effective in July 2016, and could be used to support the monitoring again in July 2017. Many people still do of NCD tax policy by adding items that are not have a clear idea of what unhealthy subject to NCD tax and potential products are, what healthy options are, substitutes into the CPI basket (if they and how to access them. Consultations are not already in the basket), which will with key agencies, key stakeholders, help the government monitor price data church leaders, and representatives of on a monthly basis. Moreover, there is a the communities should be conducted prior to imposition of new taxation policy. 109 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Transparency in use of NCD tax In response to calls from communities, More efforts are needed to create an the Government of Tonga may explore enabling environment to unite the how it can demonstrate that revenue Tongan people around the urgency to gained from NCD-related taxes is used address NCDs and to establish new to support and promote healthy lifestyles social norms and values (e.g. creating among the population, as well as the confidence that it is acceptable to improving health care services. Some make offerings with smaller meals yet of the key health promotion interventions still able to receive blessings, creating for which communities would like to the demand for “feeling healthy”). It is receive government support include: also important to explore how counselling making healthy food options cheaper, services can be made more effective to invigorating health promotion activities, strengthen one-to-one communication. supporting production of healthy food The government can also build on the products, and food preparation, etc. country’s high literacy rates. Complementary (non-tax) measures Multi-sectoral interventions are needed to address NCD burdens; fiscal/taxation Political, social and community leaders policy alone is not adequate. must be engaged to support social and Fiscal/taxation policy needs to be behavior change to address the NCD complemented by other policy crisis. From the qualitative study, it is interventions from health, education, clear that political leaders, church leaders, and community leaders are very influential, and their roles in promoting healthy lifestyles are critical. There are also superb experiences in healthy eating and healthy lifestyles that have been implemented by selected churches and communities. Their lessons should be shared widely. Good examples include the introduction of healthy food at church conferences by SDA, which encouraged the use of healthy ingredients and discouraged overeating. This is a good practice for other churches. 110 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga agriculture/fisheries sectors, with strong The National NCD Committee, its legislation, regulations, effective advisory bodies, the National Food communication and campaigns, as well as Council, and Tonga Health Promotion an enabling environment. For example, the Foundation should play an active role in government could complement taxation promoting multi-sectoral interventions interventions on Tapaka Tonga by regulating to address the NCD crisis that are to and enforcing large and clear warning be implemented by sectoral ministries. labels on packets of Tapaka Tonga, to In the past two years, the National educate the population on the harm of NCD Committee and the National Tapaka Tonga (to eliminate the general Food Council played a very limited misperception that Tapaka Tonga is not role in the design, implementation, and harmful to health), as well as providing monitoring of NCD tax and other health support or incentives (knowledge and/or promotion activities. The concerted efforts seeds) to encourage farmers to switch from by these bodies, combined by active growing tobacco to growing healthy cash implementation by relevant sectoral crops. In addition it will be important for ministries, churches and civil societies the Ministry of Health to be more involved can make a difference to Tonga and help in aligning fiscal policy with nutritional address the NCD crisis. information (nutritional value of foods, dietary requirements, etc). More policy interventions are needed to make healthy food more accessible and attractive, and also to encourage consumption of smaller portions of food. For example, fresh fish is highly desirable, but less well-off households reported that they cannot consume fresh fish on a regular basis due to the high price. The Ministry of Fisheries’ program that requires foreign fishing vessels to offload at least three tons of fish to sell to the community at T$2.5-9 per kg has been very popular. However, it seemed to benefit only communities living close to where the fish was sold, e.g. Nuku’alofa. Communities from other areas of Tongatapu and Vava’u would like to benefit from a similar program. There is the need to increase production of local vegetables and fruits as well as to raise local livestock (e.g. poultry for meat and eggs) and farm fish. Subsidies to strengthen the local food system could be considered. In addition, smaller portions of food at meal times should be encouraged and promoted. 111 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga More studies should be conducted on tasty, and popular (demand-side), home production initiatives to generate accompanied by nutritional education, evidence and promote successful cases. as well as social and behavior change For example, the government could build communication initiatives. Many on the successful Church of the Latter Tongans over the years have developed a Day Saints-supported “keyhole garden” taste for unhealthy, fatty products. Hence, initiative as an easy way to grow making healthy food more affordable vegetables that can be afforded by every alone does not guarantee that the household, even with limited land area. population will stop consuming unhealthy A keyhole garden is a raised-bed food and turn to healthy products. From system that requires minimal watering focus group discussions, communities and fertilizer due to its ability to retain have expressed that they like the taste of moisture and use compost. Its unique vegetables cooked and served at Chinese design allows farmers to irrigate their restaurants. However, they do not know crops with recycled water from how to cook vegetables and make them household washing. The keyhole structure palatable at home. Hence, communities also makes it easy for farmers to add soil expressed a desire for cooking demonstration or manure to the garden if crops need programs with recipes and techniques for additional fertilizer. A raised bed is also cooking healthy products that they can particularly effective for those who have learn to cook by themselves at home. In physical limitations, such as the elderly this way, affordable dietary change can or people with disability. There are also occur across all socio-economic groups. good examples of intercrop home gardens, as shown in Picture 11. The government Guidelines on healthy foods should should also build on communities’ requests be promoted, while cooking programs to support access to markets to stimulate should include training on dietary needs production, study experiences from poultry and portion size per age group, as well raising in other PICs (e.g. Fiji), and increase as ways to introduce simple changes community mobilization and organizations that can be incorporated into daily (e.g. farmers’ groups). cooking routines without too much effort. For example, this could include Making healthy food more affordable cooking the same meals (e.g. lu sipi), (supply side) alone is not adequate. but with less meat and more vegetables; It is important to introduce innovative washing canned fish to reduce salt ideas to make healthy food attractive, content; making home-made ice cream from frozen fruit and coconut milk; preparing breakfast with breadfruit. A focus on plant-based well-being programs could also be explored. During feasts or other events, pre-arranged packages with adequate portion sizes could be served. If such initial demand-side changes were found to be successful, at a later stage, certifications of shops/restaurants with “healthy food licenses” could be introduced on the supply side (e.g. as in Thailand or Singapore). Picture 11 : “Intercrop” home garden right behind the Central Bank of Tonga ― lettuce, onions, and kava 112 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Bibliography 113 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Cortinovis, I., Vela, V. and Ndiku, J. (1993). Construction of a socio-economic index to facilitate analysis of health in data in developing countries. Social Science and Medicine, Vol. 36. Page 1087―1097. Ebert, U. and Welsh, H. (2004). Meaningful Environmental Indices: A Social Choice Approach. Journal of Environmental Economics and Management, Vol. 47, issue 2, 270―283. Fendrich, M., M.E. Mackesy-Amiti, T.P. Johnson, A.Hubbell, and J.S. Wislar. (2005). Tobacco-Reporting Validityin an Epidemiologic Drug-Use Survey. Addictive Behavior, Vol. 30:175–181. Food and Agriculture Organisation of the United Nations and the Fiji National University Pacific Research Centre for the Prevention of Obesity and Non-communicable Diseases (C-POND). Effects of food taxation in Tonga: A Snapshot. Rome. (2017). Giovannucci E., Stampfer M.J., Colditz G.A., Manson J.E., Rosner B.A., Longnecker M.P., Speizer F.E., and Willett W.C. (1993). Recall and selection bias in reporting past alcohol consumption among breast cancer cases. Cancer Causes Control, Vol. 4(5):441―8. Global strategy on diet, physical activity and health. Geneva. World Health Organization. 2004. Gmel, G. and Daeppen, JB. (2007). Recall bias for seven-day recall measurement of alcohol consumption among emergency department patients: implications for case-crossover designs. Journal of Studies on Alcohol and Drugs, Vol 68(2):303―10. Hufanga, S., Comparative Effectiveness of Interventions to Control Non-Communicable Diseases in the Kingdom of Tonga, Draft Ph.D. Dissertation, University of New South Wales, 2018. Human Capital Index, World Bank. 2018. International Development Association program document for a proposed development policy grant to the Kingdom of Tonga for the Second Inclusive Growth Development Policy Operation, April 27, 2017. The World Bank.Kingdom of Tonga Risk Factors STEPS Report. Fiji. World Health Organization. Ministry of Health, Kingdom of Tonga. 2014. Krall, E.A., Valadian, I., Dwyer, J.T., and Dardner, J. (1989). Accuracy of recalled smoking data. American Journal of Public Health, Vol 79(2): 200―202. McLaughlin, J.K., Dietz, M.S., Mehl, E.S., and Blot, W.J. (1987). Reliability of surrogate information on cigarette smoking by type of information. American Journal Epidemiology, Vol. 126: 144―146. Ministry of Health, Kingdom of Tonga Risk Factors STEPS Report. 2014. Mumuiheata, S. (2007). Nutrient Intake in the Kingdom of Tonga. Nuku’alofa. National Strategy for Prevention and Control on Non-Communicable Diseases 2015―2020. Government of Tonga. 2016. Nicholson. C. and Snyder, W. (2016). Microeconomics: Basics, Principles and Extensions. Cengage Learning, Singapore. Obesity and Overweight Key Facts, World Health Organization. 2018. OECD. (2008). Handbook on Constructing Composite Indicators: Methodology and User Guide. Parry, C. D., Patra, J., & Rehm, J. (2011). Alcohol consumption and non-communicable diseases: epidemiology and policy implications. Addiction (Abingdon, England), 106(10), 1718―24. Patrick, D.L., A. Cheadle, D.C. Thompson, P. Diehr, T.Koepsell, and S. Kinne. (1994). The Validity of Self-Reported Smoking: A Review and Meta-Analysis. American Journal of Public Health, Vol. 84: 1086– 1093. 114 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga The Kingdom of Tonga Health System Review. Health Systems in Transition. Vol. 5, No. 6, 2015. Manila: WHO Regional Office for the Western Pacific. Sahal S. et al. (2014). Trade as a structural driver of dietary risk factors for non-communicable diseases in the Pacific: an analysis of household income and expenditure survey data. Sassi, F. et al. Equity impact of price policies to promote healthy behaviours. The Lancet Taskforce on NCDs and economics, Volume 391, Issue 10134, P2059-2070, May 19, 2018. Vyas, S. and Kumaranayake, L. (2006). Constructing socio-economic status indices: how to use principal component analysis. Health Policy and Planning, Vol. 21, issue 6, 459―468. World Health Organization. Sin Tax Expands Coverage in the Philippines, May 2015, http://www.who.int/ features/2015/ncd-philippines/en/. World Development Indicators, World Bank, 2018. Weitkunat R., Coggins, C., Sponsiello-Wang, Z., Kallischnigg, G., and Dempsey, R. (2013). Assessment of cigarette smoking in epidemiologic studies. Contributions to Tobacco Research, Vol. 25(7): 638―648. WHO Report on the Global Tobacco Epidemic, 2017: Monitoring tobacco use and prevention policies. Geneva. World Health Organization. World Bank. (2019). Confronting Illicit Tobacco Trade: a Global Review of Country Experiences: Confronting Illicit Tobacco Trade: a Global Review of Country Experiences (English). WBG Global Tobacco Control Program. Washington, D.C.: World Bank Group. 115 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Annexes 116 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Annex 1 Data 1.1 Sampling Design An initial sample of 1,400 individuals (respondents) ages 18–65 was chosen as household representatives to undertake the Tonga NCD Taxation Household Baseline Survey 2017. These households were from the two main islands, Tongatapu and Vava’u. Table 18 provides information of the total sample size and island-wise distribution. The sample was selected independently within each of the two target areas. Firstly, extremely remote areas were removed from the frame (and thus not given a chance of selection) as it was considered too costly to cover these areas. These areas only represented about 2.5 percent of the total population for Tonga, so the impact of their removal was minimal. Table 18: Total sample size of baseline data Island division Number of households Required sample Tongatapu 12,953 1,000 Vava'u 2,715 400 Total 15,668 1,400 The sampling in each area was then undertaken using a two-stage process. The first stage involved the selection of required sample from each village using Probability Proportional to Size (PPS) sampling with Number of Household as the size. For the second stage, a list of all households in every village included in the frame was provided from the Tonga National Population and Housing Census 2016. Head of Households lists for all villages were sorted by alphabetical order and then a systematic sampling was used to draw out each required sample from each village with their name, and even with a name of a household member who identified as a smoker. This gave every household equal chance of getting selected, and selected households are well representative of the general households. GPS coordinates were provided to help enumerators to locate the households. The enumerator interviewed the Head of household or someone who was capable of answering the questions. If that person did not smoke, then the name of the smoker given on the list would be called to continue the tobacco question; the same process was applicable to questions on alcohol consumption. The above methodology was used in the collection of endline data. The sample for the NCD Taxation Endline Survey 2017 was drawn from the master sample frame of Household List from the most recent Population and Housing Census, 2016. The sample was designed to cover 9 percent of the households for all Tonga excluding the Ongo Niuas and some small islands due to the cost of getting to these islands. As such, the sample size is recorded to be 1,600 with the above distribution between the Islands. Details of the number of households by Islands are shown in Table 19. 117 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 19: Total sample size of endline data Island/division Number of households Required sample Tongatapu 12,944 1,000 Vava'u 2,503 400 Ha'apai 1,179 114 'Eua 885 86 Total 17,511 1,600 1.2 Descriptive statistics of baseline data The original dataset contains 1,402 household observations. After dropping observations with missing values, 1,395 household observations could be used for subsequent analysis. There are 996 household observations from Tongatapu; the remaining 399 household observations are from Vava’u. Table 20 provides descriptive statistics on socio-economic background of the respondents. Around 65 percent of the data relate to males. Average age of respondents is 39.43. Their average monthly earnings are 996.60 Tongan Pa’anga. Average monthly earnings for males is 1,005 Tongan Pa’anga which is slightly more than females, whose earnings are approximately 980 Tongan Pa’anga per month. Household sizes of both genders are slightly above six members per household. However, there is a great variation of number of household members ranging from one to 24. When examining marital status, 70.18 percent are married and 23.58 percent are never married. Proportion of males with never-married status is higher than females by 6 percent. Island-wise descriptive statistics are shown in Table 21. 118 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 20: Descriptive statistics of baseline data Variable Male Female Overall Min Max Gender 65.73% 34.27% 100.00% - - Age 39.73 38.85 39.43 18 65 Earning 1,005.32 979.88 996.6043 0 50,000 Number of household members 6.14 6.69 6.33 1 24 Education No formal schooling 0.11% 0.00% 0.07% - - Primary school 2.18% 2.30% 2.22% - - Lower secondary 26.28% 15.69% 22.65% - - Upper secondary 54.53% 65.90% 58.42% - - Technical and vocational 11.12% 9.00% 10.39% - - Tertiary 5.23% 6.49% 5.66% - - Other 0.55% 0.63% 0.57% - - Marital status Never married 25.63% 19.67% 23.58% - - Currently married 69.57% 71.34% 70.18% - - De facto or consensual marriage 0.44% 0.42% 0.43% - - Separated 0.98% 1.46% 1.15% - - Divorced 0.44% 1.88% 0.93% - - Widowed 2.94% 5.02% 3.66% - - Other 0.00% 0.21% 0.07% - - Occupation Employer 3.16% 0.42% 2.22% - - Self-employed 13.96% 15.69% 14.55% - - Employee, working for wages (type 1) 9.60% 5.02% 8.03% - - Employee, working for wages (type 2) 24.65% 9.62% 19.50% - - Producing goods for own and/ or family 23.23% 3.77% 16.56% - - Unpaid family worker (family business) 5.56% 0.63% 3.87% - - Unpaid family worker 1.74% 12.76% 5.52% - - (help with basic household activities) Volunteer work 0.98% 0.00% 0.65% - - Student ― full time 2.18% 1.88% 2.08% - - Student ― part time 0.44% 0.63% 0.50% - - Home duties 12.43% 49.16% 25.02% - - Retired 1.53% 0.21% 1.08% - - None ― did not pursue any activity 0.44% 0.21% 0.36% - - Physically/mentally disabled 0.11% 0.00% 0.07% - - Total number of observations 917 478 1,395 119 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 21: Baseline descriptive statistics by island Tongatapu Vava’u Variable Male Female Overall Min Max Male Female Overall Min Max Gender 65.96% 34.04% 100.00% - - 65.16% 34.84% 100.00% - - Age 39.35 38.82 39.17 18 65 40.7 38.91 40.08 18 64 Earning 1,124.13 1,072.12 1,106.30 0 50,000 705.12 754.91 722.46 100 20,000 Number of household 6.31 6.89 6.5 1 23 5.73 6.2 5.89 1 24 members Education No formal schooling - - - - - 0.38% 0.00% 0.25% - - Primary school 1.98% 2.95% 2.31% - - 2.69% 0.72% 2.01% - - Lower secondary 21.92% 12.68% 18.78% - - 37.31% 23.02% 32.33% - - Upper secondary 56.77% 65.19% 59.64% - - 48.85% 67.63% 55.39% - - Technical and vocational 13.55% 11.50% 12.85% - - 5.00% 2.88% 4.26% - - Tertiary 5.02% 6.78% 5.62% - - 5.77% 5.76% 5.76% - - Other 0.76% 0.88% 0.80% - - - - - - - Marital status Never married 25.57% 22.42% 24.50% - - 25.77% 12.95% 21.30% - - Currently married 69.56% 67.55% 68.88% - - 69.62% 80.58% 73.43% - - De facto or consensual 0.61% 0.29% 0.50% - - 0.00% 0.72% 0.25% - - marriage Separated 1.07% 1.47% 1.20% - - 0.77% 1.44% 1.00% - - Divorced 0.46% 2.65% 1.20% - - 0.38% 0.00% 0.25% - - Widowed 2.74% 5.31% 3.61% - - 3.46% 4.32% 3.76% - - Other 0.00% 0.29% 0.10% - - - - - - - Occupation Employer 2.28% 0.59% 1.71% - - 5.38% 0.00% 3.51% - - Self-employed 14.61% 14.16% 14.46% - - 12.31% 19.42% 14.79% - - Employee, working for 9.59% 4.13% 7.73% - - 9.62% 7.19% 8.77% - - wages (type 1) Employee, working for 28.77% 9.44% 22.19% - - 14.23% 10.07% 12.78% - - wages (type 2) Producing goods for 17.20% 2.65% 12.25% - - 38.46% 6.47% 27.32% - - own and/ or family Unpaid family worker 3.81% 0.29% 2.61% - - 10.00% 1.44% 7.02% - - (family business) Unpaid family worker (help with basic 1.52% 0.88% 1.31% - - 2.31% 41.73% 16.04% - - household activities) Volunteer work 0.61% 0.00% 0.40% - - 1.92% 0.00% 1.25% - - Student ― full time 3.04% 2.65% 2.91% - - - - - - - Student ― part time 0.46% 0.88% 0.60% - - 0.38% 0.00% 0.25% - - Home duties 15.53% 64.01% 32.03% - - 4.62% 12.95% 7.52% - - Retired 1.98% 0.00% 1.31% - - 0.38% 0.72% 0.50% - - None ― did not pursue 0.61% 0.29% 0.50% - - - - - - - any activity Physically/mentally - - - - - 0.38% 0.00% 0.25% - - disabled Total number of 657 339 996 260 139 399 observations Source: Baseline survey data (2017). 120 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Annex 2 Data Descriptive statistics of endline data The original dataset contains 1,604 household observations. After dropping observations with missing values, 1,573 household observations can be used for subsequent analysis. It was possible that more than one member in a household was interviewed if the person smoked or drank alcohol. Thus, the actual number of persons interviewed was more than number of household observations. There are 1,475 individual observations from Tongatapu; there are 609 individual observations from Vava’u; 131 individual observations are from Ha’apai; and the remaining 86 individual observations are from Eua. Table 22 provides descriptive statistics on socio-economic background of the respondents. Around 59 percent of the data relate to males. Average age of respondents is 42.70. Their average monthly earnings are 1,285 Tongan Pa’anga. Average monthly earnings for males is 1,321 Tongan Pa’anga, which is slightly more than female earnings that are approximately 1,233 Tongan Pa’anga per month. Household sizes of both genders are slightly above six members per household. However, there is a great variation in the number of household members, ranging from 1 to 22. When examining marital status, 79.87 percent are married and 13.28 percent are never married. The proportion of males with never married status is higher than female by 1 percent. Island-wise descriptive statistics are shown in Table 23. Table 22: Descriptive statistics of endline data Variable Male Female Overall Min Max Gender 58.89% 41.11% 100.00% - - Age 43.11 42.12 42.70 18 76 Earning 1,321.12 1,233.04 1,285.11 9 200,000 Number of household members 6.0 6.4 6.19 1 22 Education No formal schooling 1.47% 0.27% 0.98% - - Primary school 0.74% 0.53% 0.65% - - Lower secondary 23.66% 24.20% 23.88% - - Upper secondary 59.02% 61.97% 60.23% - - Technical and vocational 9.67% 7.71% 8.87% - - Tertiary 5.43% 5.32% 5.39% - - Other - - - - - Marital status Never married 13.81% 12.50% 13.28% - - Currently married 82.32% 76.33% 79.87% - - De facto or consensual marriage 0.18% 0.67% 0.38% - - Separated 0.74% 1.46% 1.03% - - Divorced 0.74% 1.99% 1.25% - - Widowed 2.21% 7.05% 4.19% - - Other - - - - - 121 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 22: Descriptive statistics of endline data (continued) Variable Male Female Overall Min Max Occupation Employer 2.30% 1.20% 1.85% - - Self-employed 9.94% 13.56% 11.43% - - Employee, working for wages (type 1) 14.27% 6.38% 11.04% - - Employee, working for wages (type 2) 28.08% 10.20% 20.78% - - Producing goods for own and/ or family 25.69% 12.63% 20.30% - - Unpaid family worker (family business) 6.80% 6.50% 6.69% - - Unpaid family worker 0.55% 2.66% 1.41% - - (help with basic household activities) Volunteer work 1.01% 0.13% 0.65% - - Student ― full time 0.46% 0.27% 0.38% - - Student ― part time 0.74% 0.13% 0.49% - - Home duties 9.58% 46.01% 24.48% - - Retired 0.46% 0.27% 0.38% - - None ― did not pursue any activity - - - - - Physically/Mentally Disabled 0.09% 0.00% 0.05% - - Total number of observations 1,355 946 2,301 122 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 23: Endline descriptive statistics by island Tongatapu Vava’u Variable Male Female Overall Min Max Male Female Overall Min Max Gender 60.00% 40.00% 100.00% - - 59.61% 40.39% 100.00% - - Age 43.28 42.16 42.84 18 76 43.49 42.92 43.26 18 69 Earning 1,467.34 1,267.23 1,387.64 9 200,000 1,126.18 1,132.40 1,128.60 9 5,366 Number of household 6.23 6.84 6.47 1 22 6.07 5.9 6 1 20 members Education No formal schooling 2.28% 0.43% 1.54% - - - - - - - Primary school 0.57% 0.43% 0.51% - - 1.08% 1.13% 1.10% - - Lower secondary 20.83% 16.77% 19.21% - - 29.86% 49.15% 37.36% - - Upper secondary 57.49% 69.25% 62.18% - - 61.87% 40.68% 53.63% - - Technical and 11.70% 7.74% 10.12% - - 4.68% 4.52% 4.62% - - vocational Tertiary 7.13% 5.38% 6.43% - - 2.52% 4.52% 3.30% - - Other - - - - - - - - - - Marital status Never married 13.98% 12.04% 13.21% - - 11.51% 13.56% 12.31% - - Currently married 82.31% 76.34% 79.93% - - 84.89% 79.10% 82.64% - - De facto or 0.29% 0.22% 0.26% - - - - - - - consensual marriage Separated 0.71% 1.51% 1.03% - - 0.72% 1.69% 1.10% - - Divorced 0.86% 1.72% 1.20% - - 0.72% 0.56% 0.66% - - Widowed 1.85% 8.17% 4.37% - - 2.16% 5.08% 3.30% - - Other - - - - - - - - - - Occupation Employer 3.00% 0.86% 2.14% - - 1.08% 1.13% 1.10% - - Self-employed 9.42% 14.41% 11.41% - - 12.23% 12.99% 12.53% - - Employee, working 14.98% 4.30% 10.72% - - 11.15% 7.91% 9.89% - - for wages (type 1) Employee, working 30.81% 11.40% 23.07% - - 26.26% 7.91% 19.12% - - for wages (type 2) Producing goods for 19.69% 14.41% 17.58% - - 38.49% 3.39% 24.84% - - own and/ or family Unpaid family worker 4.99% 0.86% 3.34% - - 4.32% 0.56% 2.86% - - (family business) Unpaid family worker (help with basic 0.86% 4.09% 2.14% - - 0.00% 0.56% 0.22% - - household activities) Volunteer work 1.14% 0.00% 0.69% - - 1.08% 0.00% 0.66% - - Student ― full time 0.71% 0.22% 0.51% - - 0.00% 0.56% 0.22% - - Student ― part time 1.00% 0.00% 0.60% - - 0.00% 0.56% 0.22% - - Home duties 12.84% 49.25% 27.36% - - 4.68% 63.84% 27.69% - - Retired 0.43% 0.22% 0.34% - - 0.72% 0.56% 0.66% - - None ― did not pursue - - - - - - - - - - any activity Physically/Mentally 0.14% 0.00% 0.09% - - - - - - - Disabled Total number of 885 590 1,475 363 246 609 observations 123 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 23: Endline descriptive statistics by island (continued) Ha’apai Eua Variable Male Female Overall Min Max Male Female Overall Min Max Gender 61.07% 38.93% 100.00% - - 31.40% 68.60% 100.00% - - Age 40.86 41.02 40.92 20 65 39.07 39.32 39.24 18 65 Earning 701.88 1,101.37 857.4 100 6,000 1,366.67 1,380 1,375.81 150 3,000 Number of household 4.55 5.08 4.76 1 12 4.04 5.25 4.87 1 11 members Education No formal schooling - - - - - - - - - - Primary school 1.25% 0.00% 0.76% - - - - - - - Lower secondary 26.25% 15.69% 0.76% - - 25.93% 15.25% 18.60% - - Upper secondary 68.75% 70.59% 0.76% - - 40.74% 61.02% 54.65% - - Technical and 2.50% 0.00% 0.76% - - 29.63% 23.73% 25.58% - - vocational Tertiary 1.25% 13.73% 0.76% - - 3.70% 0.00% 1.16% - - Other - - - - - - - - - - Marital status Never married 16.25% 17.65% 16.79% - - 25.93% 8.47% 13.95% - - Currently married 77.50% 74.51% 76.34% - - 70.37% 69.49% 69.77% - - De facto or - - - - - 0.00% 6.77% 69.76% - - consensual marriage Separated - - - - - 3.70% 1.69% 2.33% - - Divorced 0.00% 3.92% 1.53% - - 0.00% 6.78% 4.65% - - Widowed 6.25% 3.92% 5.34% - - 0.00% 6.78% 4.65% - - Other - - - - - - - - - - Occupation Employer - - - - - 3.70% 5.08% 4.65% - - Self-employed 6.25% 7.84% 6.87% - - 11.11% 13.56% 12.79% - - Employee, working 20.00% 11.76% 16.79% - - 11.11% 13.56% 12.79% - - for wages (type 1) Employee, working 16.25% 11.76% 14.50% - - 11.11% 6.78% 8.14% - - for wages (type 2) Producing goods for 36.25% 15.69% 28.24% - - 18.52% 23.73% 22.09% - - own and/ or family Unpaid family worker 21.25% 52.94% 33.59% - - 37.04% 28.81% 31.40% - - (family business) Unpaid family worker (help with basic - - - - - - - - - - household activities) Volunteer work - - - - - 0.00% 1.69% 1.16% - - Student ― full time - - - - - - - - - - Student ― part time - - - - - 3.70% 0.00% 1.16% - - Home duties - - - - - 3.70% 6.78% 5.81% - - Retired - - - - - - - - - - None ― did not pursue - - - - - - - - - - any activity Physically/Mentally - - - - - - - - - - Disabled Total number of 80 51 131 27 59 86 observations 124 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Annex 3 Detailed information on the representativeness of household surveys To assess the representativeness of the data, distribution of households among the four islands, and gender distribution, had to be examined. Figure 83 provides graphic assessment of data representativeness of households. As can be seen, the actual proportion collected closely resembles the required proportion of observations of these islands as suggested by the Department of Statistics. Figure 83: Distribution of households by island Source: Endline survey data (2017). The issue of gender was explored in two steps. The first was to assess the gender distribution of household heads. Then the team examined the of gender distribution of all the respondents. Data from the Household Income and Expenditure Survey 2016 (HIES) was used as a benchmark against the endline data (see Table 24). It can be seen that proportion of household head in the HIES is 79.98 percent and proportion in the endline is 83.28 percent; proportions of female for HIES and the endline are 20.02 and 16.72, respectively. Using selected variables of interest, Table 25 illustrates how much the gender distributional differences influence the survey results. The weight assigned to males is calculated by dividing the male proportion in the HIES by the male proportion in the endline; female weight is calculated in similar fashion. Unweighted and weighted earnings differ only slightly; the same is true for the proportion of those who heard about the July 2017 food tax and biscuits price. These results should provide some confidence in using the endline data for further analysis. Table 24: Head of household distribution by gender Table 25: Comparison of unweighted and weighted results Gender HIES Endline Variable Unweighted Weighted Male 79.98 83.28 Earning 1320.129 1360.481 Heard about July Female 20.02 16.72 48.97 50.72 2017’s Food Tax Total 100.00 100.00 Biscuits price 5.39 5.90 Source: HIES (2016) and endline survey (2017). Source: HIES (2016) and endline survey (2017). 125 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Table 26 shows that the gender distributions of the first person in the household for males and females are 56.26 and 43.74 percent, respectively. According to the 2016 Tongan Census, there were 50,255 males and 50,396 females in Tonga, making the proportion approximately 50:50. If we look at gender distributions of the first person in the household from the survey, it is 56.26 for males and 43.74 for females. This may not resemble gender-wide population distribution because the first person is head of the household, thus the distribution should depart from the population at large; however, the deviation is not too significant to be of concern. When we examine gender-wide distribution of all persons in the household, the proportion of males increases to 62.18 percent, with the remaining 37.82 percent being female. At first glance, one might expect the distribution to resemble population characteristics, but the significant difference occurs because, once the first person in the household was interviewed, the interviewer asked if there were any other household member who used tobacco or alcohol. If there is such a person, the interviewer proceeded to interview the person on smoking and drinking behavior. Table 26: Gender distribution Gender First person in the household All persons in the household Male 56.26 62.18 Female 43.74 37.82 Total 100.00 100.00 Source: HIES (2016) and endline survey (2017). 126 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Annex 4 Data Analysis Construction of wealth index Data are analyzed using descriptive statistics such as means, median, standard deviation, minimum and maximum value. Analyses are done at both aggregate level, island-wise and by socio-economic background. To provide analysis beyond the mean of the data, differences in socio-economic background must be taken into account. The method employed here follows Ebert and Welsh (2004) and OECD (2008). It begins with selecting a set of socio-economic variables that would capture key characteristics of a household. We use types of wall materials, source of lighting, types of fuel used for cooking, types of toilet facility, cars, presence of refrigerator, washing machine, desktop computer, laptop computer, earning and education of head of household as a set of information to capture socio-economic status of a household. These variables are then analyzed by principle component analysis (PCA) to obtain a statistical model. In essence, PCA is a multivariate statistical method that helps to reduce the number of variables in a dataset into a smaller dimension. This is achieved by converting a set of correlated variables of interest into a set of uncorrelated components. In doing so, PCA enables researchers to determine the most appropriate weightings option for each variable; consequentially, an index that captures maximum variables can be derived. Further technical discussion of PCA and its application to wealth index construction can be found in Cortinovis, Vela and Ndiku (1993) and Vyas and Kumaranayake (2006). In the context of this study, the model contains information on relative importance of the variables mentioned above in representing a household’s socio-economic status. This model is used to predict a socio-economic score for each of the household. The next step is to calculate the socio-economic index (SI) for each household i. This is done by the formula shown below. In this formula hsi is socio-economic score for each of the household obtained from PCA exercise above. Min(hsi) and Max (hsi) represent minimum and maximum socio-economic scores respectively. This the value of SI will be between 0 to 1; higher SI score denote higher socio-economic status of the household. These values are then divided into quintiles for further data analysis. hsi -Min(hsi ) SIi = Max(hsi )-Min(hsi ) Elasticity calculation Elasticity is a measure of consumers’ responsiveness to changes product prices, including tax-induced price changes. To ensure that the measures are comparable across different product categories whose measurement units differ, elasticity values must be unit-free. The formula used to calculate the value is the following: 127 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Typically, the value of price elasticity is negative – i.e. when a consumer reduces their consumption when there is a price rise. Since tax increases product price, we can use the above formula to measure the effect of tax imposition. The assumption underlying the empirical accuracy of the elasticity estimate is that price increase is solely due to tax increase. For further discussion, refer to Nicholson and Snyder (2016). The accuracy of cigarette demand elasticity estimates depends on data quality. Several factors influence elasticity values. Some of these factors have already been explained in the previous section. In addition, factors related to socio-economic status also play a role in determining the magnitude of elasticity. To obtain reliable estimates of elasticity, one of two approaches must be adopted. The first approach is to take these variables into account; this can be achieved by using a multivariate statistical model. The second approach, which we adopted, focuses on controlling the influence of these variables. When estimating tax elasticities, we used individual-level data. While the baseline and the endline data were not designed to be longitudinal, we succeeded in creating panel data using the following method: Step 1: We created household-level wealth index. Step 2: The wealth index was divided into quintiles. Step 3: For each quintile, we match these two datasets using island, village, and head of the household name. Step 4: We calculate elasticity for individuals, using price and consumption data. Step 5: (Wealth) Quintile-wise elasticities were calculated as the average of individual elasticities of the same wealth quintile. The creating of panel data helped to ensure that the influence of socio-economic factors was minimal, thereby increasing the accuracy of elasticity estimates. When combined with our attempts to minimize recall bias, we believe that our results are valid. For the calculation of commodity elasticity, import quantities and price data before and after tax imposition were used. Table 27 provides an example of the use of import quantities and prices to estimate elasticity for SSBs. Note that all values reported in Figure 80 used this method to calculate elasticity. Prior to adopting this approach, we consulted with Tonga’s Ministry of Finance officials to ensure that there was no other factor that would affect prices of products under our interest during the period that fell within the scope of this study. Table 27: Calculation of SSB elasticity Price change Quantity change Elasticity % Change in After-tax price Before-tax price After-tax Before-tax %Change price 2.4 2 20.00 3,912 4,808 -18.6 -0.93 128 Using Taxation to Address Noncommunicable Diseases: Lessons from Tonga Annex 5 List of key policymakers and stakeholders who participated in the key informant interviews. No. Participants Government Ministers and Former Minister 1 Rev. Dr. Pohiva Tu’I’onetoa, Minister for Finance and National Planning 2 Hon. Dr. Saia Piukala, Minister for Health 3 Hon. Semisi Fakahau, Minister for Agriculture, Forestry, Food and Fisheries 4 Hon. Penisimani Fifita, Minister of Education and Training 5 Mr. Tevita Lavemaau, Member of Parliament, Former Minister of Finance and National Planning Government CEOs and Former CEO 6 Mrs. Balwyn Fa’otusia, CEO, Ministry of Finance and National Planning 7 Dr. Siale 'Akau'ola, CEO, Ministry of Health 8 Mr. Kelemete Vahe, CEO, Ministry of Revenue and Customs 9 Mr. Kulu ‘Anisi Bloomfield, former CEO, Ministry of Revenue and Customs Non-Government Organizations 10 Ms. Ofeina Filimoeaha, CEO, Tonga Health Promotion Foundation 11 Ms. Jeanie McKenzie, Tonga Health NCD Adviser 12 Mrs. Mele Amanaki, National Food Council 13 Mrs. Papiloa Foliaki, T.C.D.T (Tonga Trust) 14 Mrs. Amelia T. Hoponoa, Executive Director, Tonga Family Health Association 15 Mr. Tevita Halahala, Civil Society Government Senior Technical Staff 16 Mr. Heiloni Latu, Ministry of Revenue and Customs 17 Mr. Sione Faleafa, Ministry of Finance and National Planning 18 Ms. Sandra Fifita, Ministry of Commerce, Consumer, Trade, Innovation and Labour 19 Mr. Sioape Kupu, Ministry of Health 20 Mr. Anthony Cocker, Principal Consumer Affairs Officer Commerce Division 21 Mr. Metuisela Falesiva, Head of Food Division 22 Mr. Eva Mafi, Health Promotion Unit, Ministry of Health Church Leaders 23 Pastor Saia Vea, President, Seventh-day Adventist Church 24 Dr. Tuípulotu Katoanga, President, Church of Tonga 25 Mr. Silongoatonga Samani , Head of Welfare and Humanitarian, Tonga, Church of Jesus Christ of Latter-day Saints 26 Dr. Meleána Puloka, President, FWC Education Community Representative 27 Mr. Álotaisi Takau, District Officer, Kolofoóu Private Sector 28 Mr. Min Seng Cai, Local Tobacco Company 29 Mrs. Maliana Tohi, Molisi Supermarket Manager/ CEO 129