Objective To examine the potential impact of the goods and services tax (GST) on price, consumption and tax revenue from tobacco products in India and across states.
Methods Data on prices, tax rates and tax revenue are used to estimate a benchmark scenario prior to the GST implementation in 2017–2018. Using own-price elasticity of demand for tobacco products, we estimate the impact of GST at the state level and the aggregate state-wise impact to obtain the national level impact.
Findings We predict that the statutory GST rate of 28% plus compensation cess will increase the price of cigarettes, bidi and smokeless tobacco by 0.18%, 8.8% and 6%, respectively, and, as a result, it will reduce the weighted average consumption by 0.3%, 10% and 6% and increase tax revenue by 0.17%, 35% and 4.7%, respectively. Most states will experience a fall in tax revenue from tobacco products by more than 50% compared with the value-added tax revenue they collected pre-GST. The GST compensation cess will have to be used to compensate states’ revenue shortfalls.
Conclusions The GST has increased the complexity of the tax system for tobacco products. In particular, for cigarettes, the specific cess constitutes the majority of tax revenue, and therefore, it must be revised regularly to protect revenue and keep real prices from falling. In addition, India should reintroduce excise taxation on all tobacco products, at least in part, on top of the GST, to make them less affordable and to smooth the transition after the expiration of the compensation cess.
- low/middle income country
Statistics from Altmetric.com
Under India’s federal administration, the central government levies an excise duty on the production of tobacco while states are empowered to tax tobacco sales. Before the introduction of the goods and services tax (GST), excise duty on bidi and cigarettes was based on consumption units (generally per 1000 sticks) irrespective of their prices, equivalent to a specific tax. However, for most smokeless tobacco (SLT) products, excise duty was a presumptive tax based on the perceived production capacity of machines. In effect, this was equivalent to an ad-valorem tax which is set in per cent and based on the ex-factory value of products. For cigarettes, the excise depended on length and presence of a filter tip, resulting in six different tax tiers and rates. For bidi, the excise depended on whether the sticks were machine-made or hand-made, although more than 98% of bidi produced are hand-made.1 For SLT too, the excise varied depending on products. The pre-GST excise on all tobacco products included the basic excise duty, National Calamity Contingent Duty (NCCD) and a health cess. At the state level, a value-added tax (VAT) was imposed on the price inclusive of excise. This VAT varied between products within each state and within products across different states.
The GST implemented by India in July 2017 fundamentally reformed the indirect tax system by subsuming national excise duties, state-level VAT and several other duties into a single system. The GST council fixed a statutory (exclusive) ad-valorem GST rate of 28% on all tobacco products with an additional compensation cess on cigarettes and SLT. This cess is also applied on value added at every stage of the supply chain, along with the GST. The compensation cess was introduced with the special purpose of compensating states for potential revenue losses under the GST, if any, only for a period of 5 years. This compensation was necessary since the GST is a destination-based system which benefits consuming states, while the pre-GST system was origin-based and benefited manufacturing states. While the GST subsumes the excise, VAT and other cesses, the NCCD continues to apply on tobacco products under the GST.
Under the pre-GST regime, states’ VAT rates on tobacco products varied significantly from 0% to 65%, implying large variations in prices across states. This source of price variation is expected to vanish under the GST. Consequently, the GST will differentially affect consumption and revenue from tobacco products across states, with winners and losers depending on the difference between pre-GST and post-GST rates in each state. In addition, only one half (14%) of the GST is allocated to states (SGST) while the other half is allocated to the centre (CGST). The compensation cess and NCCD revenue are remitted to the central government. The redistribution of compensation cess revenue is not immediate as its proceeds are temporarily credited to a non-lapsable fund, until changes in state tax revenue are ascertained.
In adopting the GST, the goal of the government was to simplify and unify the complex web of central, states and local sales taxes prevailed pre-GST. Presumably, a simpler tax system would reduce administrative burden and increase consumers’ compliance. However, since the GST is an ad-valorem tax in practice, it bears the risk of increasing price variation across tobacco products and brands.
About 28.6% of persons (267 million) aged 15 years or older use tobacco in some form in India in 2017 out of which, about 100 million smoke tobacco and the rest uses SLT.2 Available estimates show that, in India, about 930 000 people die due to smoking3 and about 350 000 die due to SLT use4 every year, together accounting for about 1 280 000 deaths per year or approximately 3500 deaths per day. The total economic cost of tobacco use resulting from direct health expenditures and indirect costs from premature deaths and productivity loss is enormous, estimated to represent 1.16% of India’s gross domestic product in 2011.5 To reduce tobacco consumption, any tax reform that affects tobacco prices must simultaneously reduce their affordability.
GST being a major overhaul in the history of India’s indirect taxation, it is important to assess its effects on tobacco products. This paper estimates the impact of the GST reform on prices, consumption and tax revenue of cigarettes, bidi and SLT at the national and states levels. It focuses on predicted rather than actual outcomes in order to isolate the effects of GST from other supply and demand factors such as industry pricing strategies.
We start by simulating a baseline scenario of price, consumption and VAT revenue under the pre-GST system for financial year (FY) 2017–2018. We first compute the wholesale prices of bidi and cigarettes from their retail prices using the following formula:
where WP=whole sale price; P=retail price; VAT=exclusive VAT rate, ED=specific excise duty and R=retail margin on pretax price. Similarly, for SLT, the WP derived as:
where ER is the exclusive ad-valorem excise rate. To estimate the retail price under the GST, we apply the GST structure to the baseline WP, that is, the GST rate is applied on taxable value at every stage of the supply chain, starting with the wholesale stage, and following with the retail stage (inclusive of retail margin). Consistent with industry’s existing pricing structure,6 7 we seek a final retail margin of 10% which is approximately equivalent to applying a value of 12% for R in equations (1) and (2).
The baseline volume of consumption, and the unit price for cigarettes and SLT were obtained from Euromonitor6 and the custom report on SLT by Euromonitor for the WHO,8 respectively. Since Euromonitor data cover calendar years, while Indian financial year covers 1 April to 31 March, we use an average of 2016 and 2017 consumption and price data as our baseline for cigarettes and SLT. Baseline bidi consumption was estimated by dividing the actual excise revenue from bidi in FY 2016–2017 by the per stick excise duty rate obtained from the Ministry of Finance.9 The baseline average retail price of a pack of 25 bidi sticks was obtained from retail price data by the Labour Bureau10 for FY 2017–2018. A weighted average specific excise is estimated for FY 2017–2018 for cigarettes using consumption shares by length7 and for bidi using machine and handmade bidi market shares.1 Similarly, a specific tax equivalent excise is estimated for a pack of 10 g of SLT. The baseline weighted average VAT for FY 2017–2018 was estimated from the states’ VAT rates in FY 2016–2017 using state-wise consumption shares for each tobacco product obtained from the Global Adult Tobacco Survey (GATS) 2016–20172 as weights.
Next, we simulate the impact of the GST tax reform on prices, consumption and tax revenue, and compare it to the baseline scenario. Changes in consumption resulting from changes in taxes or retail prices depend on the price responsiveness of tobacco products, summarised into the own-price elasticity of demand. Few studies11–14 have estimated the price elasticity of tobacco products in India and there is only one study11 that provides own-price elasticities of all three products separately for India as a whole. It estimated own-price elasticity of bidi, cigarettes and leaf tobacco as −0.91,–0.34 and −0.88, respectively,11 and we used these for the present simulation. The cross-price effects are not considered since the literature on price elasticities in India finds largely insignificant cross-price effects between different tobacco products.11–14 Moreover, since we simulated only the instantaneous impact of a change of tax regime at a given point of time, ceteris paribus, using of income elasticity is unwarranted.
In order to calculate the impact of the GST at the state-level, we construct a baseline scenario of retail prices, consumption and VAT revenue from each tobacco products in each state for FY 2017–2018. Assuming that VAT is the main source of between-state price variation before GST, the baseline retail price is derived using the same national level wholesale price and state-specific VAT rates for each product. Post-GST, between-state price variation is predicted to disappear as the GST rates are uniform across states for each tobacco product. The larger the relative difference in tobacco product prices pre-GST and post-GST in a state, the larger its impact on consumption and tax revenue. State VAT revenue under the baseline are obtained using states’ consumption shares for each product, states’ VAT rates and imputed state-level prices. Assuming the same price elasticity of demand across states, we simulate the impact of the GST on consumption and tax revenue for each state by tobacco product. The weighted average percentage change in price estimated from the state level analysis is used to examine the change in national level consumption and tax revenue from each tobacco product.
Impacts at the national level
Table 1 presents the tax structures of different tobacco products before and after the GST. It also demonstrates the complexity of tobacco taxation in India before the GST with taxes varying by different product attributes and what GST does either to simplify the system or increase its complexity. The GST structure for tobacco products includes a GST rate of 28%, an NCCD applied on all tobacco products and a compensation cess applied only to cigarettes and SLT. While the GST eliminated the pre-GST distinction in excise between machine-made and hand-made bidi, the tiered system for cigarettes was retained. While the earlier system had NCCD and excise collected together as a fully specific tax for cigarettes and VAT had an ad-valorem nature, the GST subsumed excise and VAT while retaining NCCD to be collected separately, and adding a compensation cess with both ad-valorem and specific components apart from the ad-valorem GST itself. All in all, while an important goal of the GST was to simplify indirect taxation, it effectively generated additional complexity in the taxation of tobacco products, especially that of cigarettes.
One-on-one comparison of pre-GST and post-GST statutory rates in table 1 may ostensibly imply that the taxation on all tobacco products is higher post-GST than before. However, this hides that, under the pre-GST system, the VAT was levied on the retail price inclusive of excise, leading to cascading of tax (double or multiple taxation). By contrast, under the new system, the GST and ad-valorem cesses at the production stage are only levied on the ex-factory component. The NCCD is levied at a very small rate under both systems. To correctly compare the two systems, we compute the effective tax burden, or the total tax expressed as a percentage of the final retail price.
We estimate baseline annual consumption of cigarettes and bidi as 83 and 260 billion sticks, respectively, and that of SLT as 11.44 billion pouches of 10 g each, as shown in table 2. GATS2 estimated the consumption of cigarettes and bidi in 2016–2017 as more than 90 and 395 billion sticks, respectively. The discrepancy can be attributed to a number of factors including possible measurement error, tax avoidance and/or genuine tax exemption. While the only existing independent study for cigarettes in India suggests that illicit trade accounts for just 2.7%15 of the market, there is no such estimate for bidi and SLT. In the pre-GST regime, the exemption for smaller producers—protecting bidi manufacturers that produce less than two million bidi sticks annually from paying excise duties—resulted in split-offs by larger manufactures into smaller units to avoid taxation.16 Meanwhile, the proportion of unregistered SLT manufacturing has increased17 potentially resulting in more tax avoidance. Although the smaller producer exemption to bidi manufactures no longer applies under the GST, a new provision exempts businesses with aggregate turnover of less than INR 2 million (1 million in case of special category states) from GST. With an average price of INR 15 per pack of 25 sticks of bidi, this translates to approximately 3.3 million bidi sticks per manufacturer. Similarly, several SLT manufactures operating in the unregistered sector potentially benefit from this exemption, effectively reducing the tax base of bidi and SLT products under the GST on the lower-end of the supply chain, contrasting with the original intent of the new system. As these products are still subject to the GST through the wholesale distribution channel where the turnover is likely more than INR 2 million, the negative impact of the exemption on revenue might be reduced.
Table 2 compares the baseline scenario with the estimated impact of the GST reform on consumption and tax revenue from cigarettes, bidi and SLT at the national level. The post-GST consumption shown in the table is the consumption-weighted estimate from the state-level analysis. The results show that the tax burden increases marginally for cigarettes to 52.8% under the GST compared with the pre-GST level of 52.6%, and the retail price increases marginally by 0.18%. As a result, we estimate a decrease in weighted average cigarette consumption by about 0.3% (230 million sticks) and an increase in total tax revenue by about 0.17% (INR 647 million).
The GST is estimated to increase the tax burden on bidi to 22%, compared with 16% pre-GST. As a result, bidi price is expected to increase by about 8.8% (about INR 1.3 per pack of 25 sticks). Given the price elasticity of bidi demand, this results in a decrease in weighted average bidi consumption by about 10% or 26 billion sticks, while tax revenue from bidi still increases by about 35%, although the model assumes zero shifting of consumption to the tax-exempt bidi sector. For SLT products, the GST should increase the tax burden by 5% to 60% and the final retail price by about 6%, resulting in decreasing weighted average SLT consumption by about 6% and raising total tax revenue by 4.7%. The compensation cess post-GST constitutes 72% and 75% of total tax revenue from cigarettes and SLT, respectively. This, along with a similar cess raised from other commodities, represents the amount that the union government is expected to use to compensate states’ revenue shortfalls under the GST.
Impacts at the state level
Table 3 shows that the impact of the GST on consumption and tax (SGST) revenue from cigarettes varies substantially across Indian states. States are sorted based on their percentage reduction in SGST revenue in comparison to pre-GST VAT revenue without accounting for compensation cess. Rajasthan would experience the largest per cent increase in consumption of cigarettes at 8% and the largest decline of 86% in tax revenue from cigarettes compared with the baseline where states collected VAT. In all states except Manipur, the cigarette tax revenue would decline by 56% or more compared with the VAT revenue collected pre-GST because the GST directly reallocates only 14% SGST revenue from cigarettes taxation to states prior to compensation. The largest losers in tax revenue are states that used to levy a VAT on cigarettes of 30% or more. Nine states would also experience increases in cigarette consumption. In absolute terms, the decline in VAT revenue from cigarettes is expected to be the highest (INR 29.2 billion) in Uttar Pradesh, followed by West Bengal (INR 21.6 billion).
Table 4 shows the estimated impact of the GST for bidi. Rajasthan would experience the largest per cent increase of 25% in bidi consumption and a 75% decline in SGST revenue compared with the pre-GST VAT revenue prior to compensation. On the other hand, nine states would experience increase in tax revenue from bidi under the GST, five of which applied no VAT on bidi pre-GST. The largest revenue losses from bidi would occur in states where the VAT on bidi was 30% or more pre-GST. Nine states would also experience an increase in bidi consumption. In absolute terms, while three states should experience a decline in tax revenue from bidi of at least INR 1 billion or more, in West Bengal, which did not impose a VAT on bidi pre-GST, would generate INR 4.3 billion of new tax revenue from bidi.
Table 5 shows the estimated impact of the GST for SLT. Madhya Pradesh would experience a 9.3% increase in SLT consumption and 82.8% decline in SGST revenue. All states would experience a decline of at least 50% in revenue from SLT compared with the VAT revenue they would have collected without GST, prior to compensation. The largest declines in tax revenue would occur in states where the VAT on SLT was 30% or more. Four states would also experience an increase in consumption.
Although one of the main goals of the GST was to simplify the indirect tax system by subsuming local and national taxes into a single national system, we show that the reform incidentally created a new complexity in the tax structure of tobacco products, especially for cigarettes and smokeless. In addition to the GST rate, it introduced two additional cesses and kept the differentiation of tax rates across product characteristics. Although the GST rate is uniform across tobacco products, bidis are still treated differently with no cess. This added complexity is not desirable because it provides new opportunities for tax avoidance. A recent report by the World Bank was critical of the complexity of the new indirect tax system in India, noting that “currently [India] has 4 non-zero GST rates (5, 12, 18, and 28 percent) […]. Most countries around the world have a single rate of GST: 49 countries use a single rate, 28 use two rates, and only 5 countries including India use four rates”.18
This study has limitations. First, the VAT rates used for the baseline scenario in 2017–2018 are assumed to be the same as the ones prevailed in 2016–2017. Nevertheless, we do not expect this assumption to significantly affect the results as VAT rates on tobacco products rarely vary over two consecutive years. Second, in cases where the variation in prices of tobacco products between neighbouring states was large under the pre-GST system, there is a possibility that some of the observed consumption in the high-VAT jurisdiction was realised from sales in the neighbouring low-VAT jurisdiction. This situation would bias the predicted consumption changes. Third, our analysis is restricted to tax-paid consumption of cigarettes, bidi and SLT. While for cigarettes the percentage of tax-paid consumption may not be much different from the actual consumption,15 this may not be the case for bidi and SLT because a large part of their production is unorganised. Nevertheless, this should not affect the size of the percentage changes estimated in this analysis. Fourth, to keep the simulation relatively simple and transparent, we use weighted average taxes and prices for cigarettes, bidi and SLT. Nevertheless, the actual prices and tax vary significantly within each product. Our choice to show the simulated impact based on the average prices and taxes makes the analysis more transparent and less likely to be subject to measurement error. Notwithstanding these caveats, this paper provides a rigorous research-based reference both at the centre and state levels to assess the immediate impact of GST on tobacco products.
This paper is the first formal analysis of the impact of the GST on price, consumption and tax revenue from tobacco products across Indian states and territories. We show that, at the country level, the GST will most likely increase the price and reduce consumption of bidi and SLT, while its impact on cigarettes is neutral in the short term. Moreover, the GST will likely reduce the tax base of tobacco products, especially that of bidi and SLT. At the state level, tax revenue from all tobacco products before GST compensation will decrease in most states by more than 50% compared with the previous system. Under the new law, the GST Council will have to use the limited compensation cess pool to cover states’ revenue shortfalls. Nevertheless, the large revenue losses from tobacco products will likely continue to impact the budgets of several state governments beyond the 5-year period when the GST compensation cess is set to expire because the GST precludes state governments’ autonomy to affect the tax on tobacco products. States that imposed high VAT rates pre-GST will experience increased demand for tobacco products post-GST, likely resulting in an even larger toll on welfare than the direct negative impacts on states' revenue which would also affect their public health gains.
We also show that the compensation cess constitutes more than two-thirds of total post-GST tax revenue from cigarettes and SLT. The government will face the need to secure a clear strategy to replace the compensation cess after its expiration both to sustain revenue and to keep the affordability of tobacco products. To achieve this, we recommend to retain at least a portion of the excise duty on top of the present GST structure, otherwise the tax burden on tobacco products under the GST will remain far below the levels recommended by the WHO.19 Moreover, future changes in GST rates will require a consensus by the GST Council, consisting of members from all states and the centre, making GST rates practically more difficult to modify in the short term than excise duties which are determined only by the union Minister of Finance.
As bidi is the second most commonly used tobacco product in India,2 any tax policy action intended to curb tobacco consumption should treat bidi similarly with other tobacco products. Public health reasons warrant that the compensation cess currently applied to cigarettes and SLT be applied to bidi as well. Moreover, the GST provision that exempts small businesses with annual turnover below INR 2 million from GST should not be available to harmful products like tobacco, especially bidi and SLT as a significant part of the production of bidi and SLT happens in the informal sector. Eliminating this exemption will simplify the tax system and reduce distortions resulting in larger decreases in tax-paid consumption compared with tax-exempt consumption in the event of future tax increases.
In future years, the specific cess component of the GST on tobacco products can change only with a decision by the GST Council. With ongoing inflation and income growth, a constant specific cess results in the erosion of the tobacco tax in real terms, and makes tobacco products increasingly more affordable. Therefore, we encourage the government to consider provisions to not only preserve the compensation cess but also increase it periodically to prevent tobacco from becoming more affordable.
What this paper adds
What is already known on this subject
The indirect tax system in India was overhauled with the introduction of goods and services tax (GST) during July 2017. While in the previous system, tobacco products were subject to a combination of state-level value-added taxes and central excise duty, the GST subsumed all taxes into one national rate and introduced an additional compensation cess.
What important gaps in knowledge exist on this topic
There is no scientific analysis of the impact of the GST on consumption and revenue of different tobacco products at the national and state level in India, hence no assessment of its effectiveness to reduce tobacco use and raise revenue and which states will win or lose post-GST.
What this paper adds
This paper finds that the current GST rates would decrease consumption of all tobacco products, although marginally for cigarettes, and the GST structure will make tobacco taxation more complex, especially for cigarettes. Without considering GST compensation cess, tax revenue from all tobacco products will decrease by more than 50% for most states under GST, compared with the VAT revenue they collected before. The GST will also likely reduce the tax base on tobacco products, especially, that of bidi and SLT, an outcome opposite to what the GST reform originally intended.
Contributors RMJ conceived the research idea. RMJ and ED developed the simulation model, methods and analysed the data. RMJ wrote the first draft. All three authors contributed to the interpretation of results, reviewing different drafts and finalising the manuscript.
Funding The Campaign for Tobacco Free Kids.
Competing interests None declared.
Patient consent Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.