Background

Smoking and tobacco taxes

An estimated 680 million smokers, more than half of the world’s smoking-addicted population, resides in the Asia-Pacific region (incl. Western Pacific and Southeast Asia) [1, 2]. Studies by the World Health Organization (WHO) suggest that 80% of the world’s smokers live in low-income and middle-income countries and that more than 80% of those dying from lung cancer by 2030 will be from those countries [3]. Low and middle-income countries, including those in the Asia-Pacific region, are comparatively more affected by cigarette addiction than other countries.

In 2012, the Asian Development Bank estimated that, in the absence of intervention, smoking would eventually kill about 267 million current and future cigarette smokers in China, India, the Philippines, Thailand, and Vietnam. The study concluded that tax-induced price increases would effectively reduce the number of smokers and incidences of smoking-related deaths, while generating substantial additional tobacco tax revenues [4].

In most Asia-Pacific countries, tobacco pricing is subject to specific excise taxes, whereas ad valorem taxes are less common. In an increasing number of cases, specific excise as well as ad valorem taxes (mixed excise) are levied (Table 1). A specific excise tax is levied on quantity, whereas an ad valorem excise tax is levied as a percentage of the value of tobacco products. Although mixed excise taxation is considered the most effective in reducing nicotine consumption, the predominant taxation regime varies by region and is mostly determined by tobacco industry characteristics, public awareness and socio-political considerations [5]. In recent years, mixed excise taxation has been the preferred form of taxation in Europe and has gained popularity in Africa and Southeast Asia, while countries in the Western Pacific maintain a preference for specific excise tax only [6].

Table 1 Cigarette excise taxes structure and cigarette consumption, and retail prices from 2006 to 2015 in 22 Asia- Pacific countries
Full size table

Global tobacco consumption and taxation policies have been strongly influenced by the WHO Framework Convention on Tobacco Control (WHO FCTC). The framework went into effect in 2005 and aims at the reduction of global tobacco use through price (tax) and non-price related measures as well as efforts to curb illicit tobacco trade. Mechanisms for scientific and technical cooperation and exchange of information assist participating parties to address tobacco control issues more effectively than previous global WHO tobacco control initiatives [6]. Since the signing of the WHO framework convention (FCTC), low- and middle-income countries in the Asia-Pacific region have substantially increased tobacco taxes. Nevertheless, tobacco taxation has remained below the WHO FCTC declared best-practice tax rate of over 75% of retail price (Table 1) [7].

Price elasticities and cigarette demand

Cigarette price elasticity in low-income or middle-income countries usually ranges between − 0.5 to − 1.05 [8,9,10,11]. However, elasticity is reportedly much lower (− 0.25 to − 0.5) in high-income countries, where cigarette affordability tends to be comparatively higher [8, 12, 13]. As for Asia-Pacific countries, estimates of cigarette price elasticity are rarely reported since research has so far focused on a restricted number of countries. That is, most studies have been cross-sectional and time series analyses on China, Bangladesh, India, Indonesia, Myanmar, Nepal, Sri Lanka, and Thailand, report price elasticity figures for China between − 0.01 to − 0.84 [14,15,16,17,18,19,20,21,22], for Bangladesh between − 0.26 to − 0.49 [23,24,25], for India between 0.15 to − 1.021 [26,27,28], for Indonesia at − 0.62 [29], for Myanmar at − 0.34 [28], for Nepal at − 0.88 [10], for Sri Lanka between − 0.29 to − 0.68 [30], and for Thailand between − 0.09 to − 0.67 [31, 32].

The goals of this study

The main contribution of this study is to estimate the price elasticity of cigarette demand in low-income and middle-income countries of the Asia-Pacific region applying panel data analysis. If compared with time-series or cross-sectional studies, panel data analyses allow researchers to control for unobservables that threaten causal inference in observational studies and offer more opportunities to explore patterns of causal relationships over longer time spans [33, 34]. Furthermore, potential effects of excise tax increases on cigarette consumption, tax revenues, and the death toll of smoking are analysed by administering hypothetical price increases based on figures derived from maximum and mean increments in cigarette prices during 1999–2015. The findings of this study may serve as an important reference for health management authorities in Asia-Pacific countries to revise tobacco prevention and control policies.

Methods

Study design and data

In this study, data of 22 Asia-Pacific countries were collected to construct a cigarette demand structure model. Kiribati, North Korea, Nauru, Timor-Leste, and the Marshall Islands were excluded from the study because of lack of data. One dependent variable and three independent variables were considered. Per capita cigarette consumption for those aged 15 and over was chosen as the dependent variable. Independent variables comprised cigarette prices, gross national income (GNI) per capita, and current status of WHO FCTC ratification.

Data regarding cigarette consumption and prices were extracted from the 1999–2015 Euromonitor International market research database [35]. Euromonitor International is recognized as a leading independent provider of global business intelligence, specializing in creating worldwide data and analysis on consumer products and services. Consumption of cigarette products was calculated based on annual cigarette consumption per capita for those aged 15 and over. The retail price for a pack of cigarettes in each country was calculated by dividing cigarette sales revenues by cigarette consumption, which was further deflated using consumer price indexes [36].

The GNI per capita data were accessed from the World Bank’s database, converted to US dollars using the World Bank Atlas method [36], divided by the midyear population, and deflated based on consumer price indexes. The current status of WHO FCTC ratification provides information on whether individual countries have ratified the convention. Countries that have already ratified the convention were given a 1 value. Data regarding the ratification of the WHO FCTC were obtained from the 2015 WHO report on global tobacco epidemic [7].

Data characteristics

Table 1 lists all variables used in the analysis and visualizes their characteristics. The structural composition of excise taxes on tobacco products in the observed countries is also shown in Table 1. Total excise taxes comprise specific excise, ad valorem excise, or mixed excise. Specific excise is the most common form. According to the 2015 WHO report on global prevalence of tobacco products, the specific excise tax proportion of retail price in 2014 was the highest in the Philippines (63.55%), followed by Sri Lanka (59.15%) and Tonga (58.82%), whereas specific excise taxes accounted for less than 50% in the remaining countries. High levels of ad valorem taxes were levied on tobacco products in Thailand (63.72), Bangladesh (61%), Myanmar (50%), Vietnam (32.5%), and China (29.3%). Total taxes were the highest in Thailand (66.58), followed by the Philippines (63.55%), Sri Lanka (63.06) and Bangladesh (61%). The Maldives did not levy taxes on cigarettes. Low total tax rates were also in evidence in Laos (7.65%), Cambodia (13.15%), and Nepal (16.29%).

In 2015, per capita cigarette consumption in the 22 Asia-Pacific countries for those aged 15 years and over was the highest in China at 95.94 packs, followed by Indonesia (50.52 packs), Vietnam (47.97 packs), and the Philippines (42.12 packs), while consumption in the other Asia-Pacific countries was below 40 packs (Table 1). With the exceptions of Indonesia, Mongolia, China, and Cambodia where cigarette consumption showed a rising trend, Indonesia experienced the highest growth (51.6%), followed by Mongolia (50%) and China (28.7%). Consumption in other countries assumed a downward trend with the Solomon Islands exhibiting the greatest decrease (41.5%), followed by Malaysia (35%).

According to WHO estimates, the smoking prevalence rate of populations aged 15 or older in the 22 observed low-income and middle-income countries is below 40%. It is the highest in Samoa at 37.3%, followed by Tonga (35.9%), Laos (35.3%), China (30.5%), Bhutan (28.8%), and Nepal (27.1%). India (13.4%), Sri Lanka (12.8%), and Fiji (10.9%) exhibit the lowest prevalence rates.

As shown in Table 1, in 2015, the average real retail price for a pack of cigarettes was the highest in the Solomon Islands at US$3.67 per pack, followed by Vanuatu (US$3.53). Considering the fluctuation of real retail prices of cigarettes between 1999 and 2015, cigarette prices across the Asia-Pacific region generally showed a rising trend, with China experiencing the greatest increase (150.4%), followed by Bhutan (62%), the Philippines (49.8%) and the Maldives (48.1%), while cigarette prices in Myanmar and Vietnam decreased by 48.3%, and 29.7%, respectively (Table 1).

Empirical specification and analysis

To calculate cigarette price elasticity, a cigarette demand structure model was constructed using cigarette consumption as the dependent variable and cigarette price, gross national income (GNI), and the ratification of WHO FCTC as explanatory variables.

To estimate price elasticities of demand for cigarettes, we have applied a conventional demand model with a linear equation in this study. We assume

$$\ln {C}_{it}={\beta}_{1i}+{\beta}_2\ln {P}_{it}+{\beta}_3\ln {GNI}_{it}+{\beta}_4{FCTC}_{it}+{\varepsilon}_{it}$$
(1)

where Cit is the annual cigarette consumption per capita in the population aged 15 and older in country i in year t (1999–2015), β1i is the intercept for country i, Pit is the cigarette real retail price per pack of 20 cigarettes in country i in year t, GNIit is per capita gross national income in country i in year t, and FCTCit is a dummy variable to describe the state of ratification of the WHO FCTC in country i in year t.

Endogeneity must be considered for the regression analysis to avoid biased estimates. Among the regressors used in our analysis, cigarette price has been reported as the most likely sources of possible endogeneity in studies on cigarette consumption [37, 38]. We addressed this issue by using cigarette price and consumer price indexes in periods t–1 as instruments for cigarette price. A weak identification test (Kleibergen-Paap Wald F) was carried out to verify instrument relevance [39].

A Hausman test was applied to determine which model should be used for the equation estimation. A rejection of the test is taken to mean that the key random-effects assumption is false and in such cases the fixed-effects estimates should be used [40].

To determine the effects of cigarette price increases on cigarette consumption, cigarette consumption in 2015 was set as the baseline for this study. Maximum and mean increments in cigarette prices during 1999–2015 were applied to simulate changes in future cigarette consumption based on the cigarette price elasticity estimated in this study. Changes in tobacco tax revenues were calculated based on changes in consumption due to price increases. Percentages of price increases were calculated using the yearly mean and maximum price increases between 1999 and 2015.

Previous research has shown that cigarette price elasticity is likely to be affected by income levels [20, 41]. To account for this income threshold effect, we performed our analysis using clusters of countries with different income levels. That is, the observed 22 countries were first grouped into two clusters according to World Bank classification: (a) low- and lower-middle income countries as well as (b) upper-middle income countries. The latter group of countries was then divided into two clusters and the first into four. Gross national income per capita data for the year 2015 (Atlas method) were used in the clustering and consideration was given to obtain clusters of approximately equal size and significantly different GNI values (Table 1).

The number of averted smoking-attributable deaths (SADs) derived from the simulated impact of price increments on the reduction in smokers and was adjusted for the fact that smoking cessation still carries considerable risks of early death [42]. The applied mortality adjustment factors were calculated for each country surveyed, assuming that 95%, 75%, 70%, 50% and 10% of those who ceased smoking when aged 15 to 29, 30 to 39, 40 to 49, 50 to 59 and at least 60 years, respectively, would remain unaffected by their previous smoking habits. Data on population stratification were extracted from the World Bank database [36].

Results

Regression results

As price elasticities may change as a result of changes in income levels, we performed six panel regressions to investigate the effect of cigarette prices on cigarette consumption. Results of the administered Hausman test t showed that the models were statistically significant at the 5% level for all six samples, indicating that the fixed effects model should be administered for these samples (Table 2). We thus applied the fixed effects model for all six samples.

Table 2 Results of fixed effect regression models of panel data, 22 Asia- Pacific countries (1999–2015)
Full size table

Instrument relevance (Kleibergen-Paap Wald F) was tested for the instrumental variables used in the regression to determine whether the instruments are invalid. The results for the lagged (t-1) cigarette price and consumer price indexes as the instrumental variables showed that the instrument is not weak for the majority of income clusters (see Table 2). Based on the results, we decided to use the fixed effects model with cigarette price as the instrument variable in our six panel regressions. Moreover, fixed effects estimation was paired with cluster-robust variance estimation as to account for heteroskedasticity and un-modeled dependence among the errors [43].

Elasticity estimates

A conventional regression model was used in this study to estimate the elasticity of demand for cigarettes. The cigarette consumption, cigarette price, and income variables were all logarithmically transformed to estimate elasticity. There were differences in the cigarette price elasticity of each income group in the Asia-Pacific region (Table 2). When GNI per capita was higher than US$6000 (Group 6), cigarette price elasticity was the highest at − 1.304 and income elasticity 0.769. When GNI per capita was between US$4000 and US$5999 (Group 5), cigarette price elasticity reached − 0.614 and income elasticity, 0.327. Cigarette price elasticity for countries with a GNI per capita lower than US$1000 was at − 0.037, the lowest value among the six income groups, and income elasticity at − 0.161. In addition, previous ratification of the WHO FCTC showed a negative and statistically significant impact on cigarette consumption.

The effects of cigarette prices on cigarette consumption, tobacco tax revenue, and smoking-related deaths

To determine the effects of tax-induced cigarette price increases on cigarette consumption, cigarette prices and consumption levels of 2015 were set as the baseline for this study. The maximum and mean annual increments in cigarette prices during 1999–2015 were used to simulate changes in future cigarette consumption based on price elasticity estimated in this study. In both scenarios, increases in cigarette prices (mean and maximum) reduced cigarette consumption the most in China (price mean and max: 9.45% and 50.96%; consumption mean and max: 12.32% and 66.45%), followed by Malaysia (price mean and max: 7.68% and 17.04%; consumption mean and max: 10.01% and 22.22%). The other countries with a large reduction in cigarette consumption were the Maldives (mean: 5.53%, max: 21.3%), Sri Lanka (mean: 4.95%, max: 9.76%), and Indonesia (mean: 4.72%, max: 13.06%). The simulation result also suggests that with an average annual cigarette price increase of 9.51% during the observed period of 1999–2015, the average annual cigarette consumption would decrease by 3.56% in the 22 Asia-Pacific countries (Table 3).

Table 3 Impact of real retail cigarette price increases in 22 Asia- Pacific countries between 1999 and 2015 on cigarette consumption per capita, tax revenue, reduction in no. of smokers and reduction in smoking-attributable deaths
Full size table

Tobacco tax revenue in 2015 was used as the baseline to predict future effects of mean changes in cigarette prices on tobacco tax revenue (Table 3). The simulation result shows that the average annual tobacco tax revenue would increase by 16.2%. Laos had the highest percentage increase in tobacco tax revenue (mean and max increase: 112.45% and 262.11%), followed by Solomon Islands (mean and max increase: 65.09% and 156.27%), Nepal (mean and max increase: 71.8% and 186.95%) and Cambodia (mean and max increase: 48.77% and 198.71%).

Results of the simulation also showed that excise tax increases could potentially avert a total of 17.96 million smoking-related deaths in low and middle-income countries of the Asia-Pacific region. Specifically, about 16 million deaths could be avoided in China; 833,014 in Indonesia; 612,049 in India; 160,999 in Malaysia; 106,851 in Thailand; and 70,477 in the Philippines (Table 3).

Discussion

Previous studies have emphasized the effectiveness of tax-induced price increases to curb tobacco consumption and their subsequent positive impact on public health and finances [44]. Notwithstanding, regional socio-economic and cultural variations may produce divergent results. This study found that among low-income as well as in middle-income countries in the Asia Pacific region, countries with a per capita GNI above US$6000 exhibited the highest cigarette price elasticity. That is, most low- and middle-income countries in the region showed considerably lower elasticity figures, contradicting previous research findings, suggesting that price elasticity in less developed countries would be higher than in advanced economies [44]. Industrial counter measures as well as public policy deficiencies may explain the differences in price elasticity estimates. That is, the tobacco industry reportedly adopts a low-price strategy to mitigate tax-induced effects on cigarette prices [45]. For example, in countries such as Bangladesh, Myanmar and Vietnam real retail prices of cigarettes haven fallen over the years despite higher tobacco taxes (Table 1). Moreover, previous studies have highlighted the necessity of implementing other anti-smoking measures outlined in the WHO FCTC, such as cessation programs, bans on advertisements, and improving public awareness, in combination with higher tobacco taxation measures [7, 13]. According to WHO reports, however, non-tax related tobacco control measures are generally inadequately implemented in less developed countries, thus reducing the effects of excise taxation policies [6].

According to our simulation, cigarette consumption would decrease the most in China, where low cigarette prices have contributed to high levels of smoking prevalence (30%) with consumption amounting to one-third of global cigarette demand [7]. Among the observed Asia-Pacific countries, Indonesia, the Maldives, Sri Lanka, Mongolia and Malaysia would also be significantly affected by additional excise taxation policies. Specifically, excise taxation would not only reduce cigarette consumption, but would also lead to additional tax revenues, which could be utilized to enforce existing anti-smoking policies as well as to finance future policy instruments. Currently, a total of 10 out of the 22 countries observed in the study have taxation levels below 40%, while no taxes are levied on tobacco in the Maldives. That is, tax rates are far below the best-practice taxation rate of over 75% suggested by the World Health Organisation (WHO). Moreover, economic growth in low- and middle-income countries has made tobacco affordable to increasing numbers of consumers. The proportion of excise taxes on cigarette retail prices would thus have to be increased accordingly to compensate for increases in purchasing power.

In general, higher tobacco taxation in countries with low (or no) cigarette taxes and high smoking prevalence reportedly has a large impact on cigarette consumption. Less developed economies levying ad valorem excise on tobacco, such as Cambodia, Bangladesh, Myanmar and Vietnam, could however introduce (additional) specific excise taxes to obtain a greater impact on cigarette prices and to mitigate the impact tobacco companies may have in attempts to counteract ad valorem taxes. Currently, smoking prevalence in these countries may exhibit limited negative effects on public health and finances, but restricted state capacities to accumulate necessary funds to implement anti-smoking measures are likely to constitute considerable burdens on national health care systems in the future. Adopting specific excise taxation on tobacco as well as establishing externally-funded anti-smoking agencies may be beneficial to alleviate the negative effects of smoking on public health.

In our study, a statistically significant relationship between the ratification of the WHO FCTC and subsequent reductions in cigarette consumption could be established for all the six income clusters, except for the income group with the lowest GNI per capita (Cambodia, Nepal, Bangladesh, and Myanmar). Moreover, previous research on cigarette price elasticity and taxation effects in the Asia-Pacific applied cross-sectional methods to investigate individual countries, whereas transnational data were analysed in this study. Integrated analysis of transnational data may, however, lead to incorrect inferences, because different countries have different cigarette consumption structures. Public health authorities in the Asia-Pacific region are thus advised to establish cigarette consumption databases and market-monitoring mechanisms to facilitate long-range tracking and analysis. Moreover, demographic factors and other effects of tax-induced price increases, such as brand switching, consumption of inferior tobacco products, and illicit trade as well as their impact on consumption were not addressed in this study and should be discussed in future research.

Conclusion

This study estimated that price elasticities of each income group in the Asia-Pacific region ranged from − 0.037 to − 1.304. Countries with a per capita GNI above US$6000 (China and Malaysia) exhibited the highest cigarette price elasticity. That is, most low income as well as middle-income countries in the region showed considerably lower elasticity figures during the study period. Higher elasticity is likely to be obtained by introducing/increasing specific excise taxes and concerted efforts to implement non-price (tax) related WHO FCTC measures, such as cessation programs, bans on advertisement, and improving public awareness. The subsequent increase in tobacco tax revenues would also be instrumental in covering expenditures related to such tobacco prevention and control programs.