On 26 May 2016 the New Zealand Government announced it would continue its programme of annual 10% tobacco tax increases from 2017 to 2020 inclusive. Since 2010, the Government has increased tobacco tax by 10% each year, albeit with a higher increase in 2010 (all above a routine annual inflation adjustment). Tobacco tax is the single most effective tool used by governments to prevent young people from taking up smoking and reducing current smoking rates, while simultaneously benefiting population health and generating large savings on healthcare spending.1-3 Given this background we aimed to project future smoking prevalence for M1ori and non-M1ori men and women under the four additional years of tobacco tax compared to no continuation of this programme (ie, tax increases to January 2016 only). This letter extends previous work reported in an online blog, that projected smoking prevalence under the scenario of tax increases until 2020 for specific ethnic groups only.4 In addition, in this letter, we also aimed to discuss how such a programme could be enhanced by using collected tobacco tax revenue to fund tobacco control activities, and the need for the NZ Government to explore more novel (and substantive) measures to increase the chances of achieving its Smokefree Nation 2025 goal. Methods We used a peer-reviewed dynamic tobacco forecasting model5,6 that itself was initially derived from a published Australian model.7 This model was used to project future smoking prevalence for a continuation of annual 10% tobacco tax increases until 2020 compared to the business-as-usual trends (albeit with the tax increases that have occurred to January 2016). Briefly, this model is built in Microsoft Excel and includes a base and forecasting model using demographic and mortality data for New Zealand. In the base model recent annual trends in smoking uptake and cessation by sex, age and ethnicity were solved by using smoking prevalence data from the 2006 and 2013 censuses. To project future smoking prevalence in New Zealand by sex and ethnicity under a series of annual 10% tax increases, these baseline trends in prevalence were tweaked by using age group-specific tobacco price elasticities (ie, responsiveness to tax increases as observed in the past). We used the following age group-specific tobacco price elasticities derived from previous BODE3 tobacco tax modelling work: -0.38 (15-20 years), -0.29 (21-24 year olds), -0.19 (25-34 years), and -0.10 (35+ years).2 The applied tobacco price elasticities were scaled up by 20% for M1ori as done in previous BODE3 tobacco tax modelling work.1 There is some international evidence suggesting greater tobacco price elasticities among low-income populations,3 and as such one would expect higher price elasticities among M1ori. While there is no direct evidence of greater responsiveness to tobacco price increases among M1ori, there is some indirect evidence from an experimental study,8 and some parallel evidence from a study suggesting greater sensitivity to food pricing policies among M1ori.9 Results Our model projects that a continuation of 10% tax increases from 2017 to 2020 inclusive will see daily adult smoking prevalence rates reduce to 20.7% and 22.1% for M1ori men and women and to 10.2% and 7.7% for non-M1ori men and women by 2020 (see Figures 1 and 2). These compare to projected smoking rates of 22.0%, 23.5%, 10.6% and 8.0% for M1ori and non-M1ori men and women respectively if such a programme had not continued beyond January 2016. Assuming a continuation of baseline trends in smoking uptake and cessation after 2020 was projected to produce further reductions in smoking rates to 17.0% and 17.6% for M1ori men and women by 2025 and to 8.3% and 6.1% for non-M1ori respectively. Furthermore, the additional four rounds of tax increases have the potential to reduce the absolute ethnic gap in smoking prevalence observed in New Zealand by nearly one percentage point in 2025 (eg, from 9.6% to 8.7% for men and from 12.4% to 11.5% for women). Figure 1: Daily adult smoking prevalence projections for M1ori and non-M1ori men under two tobacco taxation scenarios (annual tobacco tax increases until January 2016 only and also ending in January 2020). Figure 2: Daily adult smoking prevalence projections for M1ori and non-M1ori women under two tobacco taxation scenarios (annual tobacco tax increases until January 2016 only and also ending in January 2020). Discussion A continuation of the programme of 10% annual tobacco tax increases was predicted to see tobacco smoking rates reduce further, but not sufficiently to achieve a below 5% smoking prevalence for M1ori or non-M1ori men and women by 2025. Yet there is uncertainty around the stability of tobacco prevalence elasticities into the future as they may increase with tobacco reaching higher prices, or increase with changes in access to nicotine-containing electronic cigarettes (both of which would probably favour larger smoking reductions than we project). There may also be various tipping points for higher quit rates as smoking becomes more denormalised in New Zealand society or psychologically significant prices are achieved (eg, the$30 per pack threshold). Conversely, others have argued that the effect of tobacco taxation may be undermined by a growing illicit tobacco market. But in our past modelling we have shown that even a substantial growth in the illicit market had little impact on overall population-wide prevalence of tobacco smoking.2 While a continuation of the 10% annual tobacco taxation programme is a solid step forward towards the NZ Governments Smokefree 2025 goal, such a programme could be enhanced by dedicating some of the collected tobacco tax revenue to fund tobacco control activities. This revenue could particularly fund additional cessation support services and mass media campaigns targeting low-income smokers who may be disproportionally affected by tax rises if they continue to smoke. Earmarking tobacco tax for tobacco control is already common practice in various jurisdictions overseas (eg, Iceland, California, Switzerland, Vietnam).10,11 Indeed, it seems unethical for the New Zealand Government to not provide smokers with more motivation to quit and more direct support to quit when it is dealing with a highly addictive substance. (The argument about smokers paying their way or not is not directly relevant when a democratically-elected government has opted for a Smokefree Nation goal. Also the issues are very complex and not all are easily quantifiable: health costs from smokers and non-smokers harmed by second-hand smoke (SHS), superannuation costs, nuisance impacts from SHS and environmental impacts such as tobacco-related litter and fires.) Also this and past New Zealand modelling work2,6,12 strongly suggest that the need for further intensification of existing tobacco control measures or completely novel measures remains. More incremental measures such as the proposed law on plain packaging, expansion of smokefree areas (including a law for smokefree cars) and more mass media campaigns (especially targeted at M1ori and Pacific audience) remain worthwhile. Novel measures, and potentially more substantive in terms of reducing smoking, that the New Zealand Government could explore include the tobacco-free generation idea (thereby essentially preventing new generations from becoming addicted to tobacco13) or a substantive reduction in the number of tobacco retail outlets (which is currently estimated at around 6,000 outlets12).

On 26 May 2016 the New Zealand Government announced it would continue its programme of annual 10% tobacco tax increases from 2017 to 2020 inclusive. Since 2010, the Government has increased tobacco tax by 10% each year, albeit with a higher increase in 2010 (all above a routine annual inflation adjustment). Tobacco tax is the single most effective tool used by governments to prevent young people from taking up smoking and reducing current smoking rates, while simultaneously benefiting population health and generating large savings on healthcare spending.1-3 Given this background we aimed to project future smoking prevalence for M1ori and non-M1ori men and women under the four additional years of tobacco tax compared to no continuation of this programme (ie, tax increases to January 2016 only). This letter extends previous work reported in an online blog, that projected smoking prevalence under the scenario of tax increases until 2020 for specific ethnic groups only.4 In addition, in this letter, we also aimed to discuss how such a programme could be enhanced by using collected tobacco tax revenue to fund tobacco control activities, and the need for the NZ Government to explore more novel (and substantive) measures to increase the chances of achieving its Smokefree Nation 2025 goal. Methods We used a peer-reviewed dynamic tobacco forecasting model5,6 that itself was initially derived from a published Australian model.7 This model was used to project future smoking prevalence for a continuation of annual 10% tobacco tax increases until 2020 compared to the business-as-usual trends (albeit with the tax increases that have occurred to January 2016). Briefly, this model is built in Microsoft Excel and includes a base and forecasting model using demographic and mortality data for New Zealand. In the base model recent annual trends in smoking uptake and cessation by sex, age and ethnicity were solved by using smoking prevalence data from the 2006 and 2013 censuses. To project future smoking prevalence in New Zealand by sex and ethnicity under a series of annual 10% tax increases, these baseline trends in prevalence were tweaked by using age group-specific tobacco price elasticities (ie, responsiveness to tax increases as observed in the past). We used the following age group-specific tobacco price elasticities derived from previous BODE3 tobacco tax modelling work: -0.38 (15-20 years), -0.29 (21-24 year olds), -0.19 (25-34 years), and -0.10 (35+ years).2 The applied tobacco price elasticities were scaled up by 20% for M1ori as done in previous BODE3 tobacco tax modelling work.1 There is some international evidence suggesting greater tobacco price elasticities among low-income populations,3 and as such one would expect higher price elasticities among M1ori. While there is no direct evidence of greater responsiveness to tobacco price increases among M1ori, there is some indirect evidence from an experimental study,8 and some parallel evidence from a study suggesting greater sensitivity to food pricing policies among M1ori.9 Results Our model projects that a continuation of 10% tax increases from 2017 to 2020 inclusive will see daily adult smoking prevalence rates reduce to 20.7% and 22.1% for M1ori men and women and to 10.2% and 7.7% for non-M1ori men and women by 2020 (see Figures 1 and 2). These compare to projected smoking rates of 22.0%, 23.5%, 10.6% and 8.0% for M1ori and non-M1ori men and women respectively if such a programme had not continued beyond January 2016. Assuming a continuation of baseline trends in smoking uptake and cessation after 2020 was projected to produce further reductions in smoking rates to 17.0% and 17.6% for M1ori men and women by 2025 and to 8.3% and 6.1% for non-M1ori respectively. Furthermore, the additional four rounds of tax increases have the potential to reduce the absolute ethnic gap in smoking prevalence observed in New Zealand by nearly one percentage point in 2025 (eg, from 9.6% to 8.7% for men and from 12.4% to 11.5% for women). Figure 1: Daily adult smoking prevalence projections for M1ori and non-M1ori men under two tobacco taxation scenarios (annual tobacco tax increases until January 2016 only and also ending in January 2020). Figure 2: Daily adult smoking prevalence projections for M1ori and non-M1ori women under two tobacco taxation scenarios (annual tobacco tax increases until January 2016 only and also ending in January 2020). Discussion A continuation of the programme of 10% annual tobacco tax increases was predicted to see tobacco smoking rates reduce further, but not sufficiently to achieve a below 5% smoking prevalence for M1ori or non-M1ori men and women by 2025. Yet there is uncertainty around the stability of tobacco prevalence elasticities into the future as they may increase with tobacco reaching higher prices, or increase with changes in access to nicotine-containing electronic cigarettes (both of which would probably favour larger smoking reductions than we project). There may also be various tipping points for higher quit rates as smoking becomes more denormalised in New Zealand society or psychologically significant prices are achieved (eg, the$30 per pack threshold). Conversely, others have argued that the effect of tobacco taxation may be undermined by a growing illicit tobacco market. But in our past modelling we have shown that even a substantial growth in the illicit market had little impact on overall population-wide prevalence of tobacco smoking.2 While a continuation of the 10% annual tobacco taxation programme is a solid step forward towards the NZ Governments Smokefree 2025 goal, such a programme could be enhanced by dedicating some of the collected tobacco tax revenue to fund tobacco control activities. This revenue could particularly fund additional cessation support services and mass media campaigns targeting low-income smokers who may be disproportionally affected by tax rises if they continue to smoke. Earmarking tobacco tax for tobacco control is already common practice in various jurisdictions overseas (eg, Iceland, California, Switzerland, Vietnam).10,11 Indeed, it seems unethical for the New Zealand Government to not provide smokers with more motivation to quit and more direct support to quit when it is dealing with a highly addictive substance. (The argument about smokers paying their way or not is not directly relevant when a democratically-elected government has opted for a Smokefree Nation goal. Also the issues are very complex and not all are easily quantifiable: health costs from smokers and non-smokers harmed by second-hand smoke (SHS), superannuation costs, nuisance impacts from SHS and environmental impacts such as tobacco-related litter and fires.) Also this and past New Zealand modelling work2,6,12 strongly suggest that the need for further intensification of existing tobacco control measures or completely novel measures remains. More incremental measures such as the proposed law on plain packaging, expansion of smokefree areas (including a law for smokefree cars) and more mass media campaigns (especially targeted at M1ori and Pacific audience) remain worthwhile. Novel measures, and potentially more substantive in terms of reducing smoking, that the New Zealand Government could explore include the tobacco-free generation idea (thereby essentially preventing new generations from becoming addicted to tobacco13) or a substantive reduction in the number of tobacco retail outlets (which is currently estimated at around 6,000 outlets12).