75

Chapter IV
The political economy of tobacco taxation

Excise taxes are an effective tool for generating higher revenues. In recent
years, in addition to satisfying revenue needs, an increasing number of govern-
ments have used tobacco tax increases in order to reduce the health and economic
burden of tobacco use. Studies have shown that tobacco taxes are the most cost
effective way to reduce tobacco consumption. Implementation of a package of
price and non-price policies (e.g. banning smoking in public places, banning
advertising etc.) is also highly cost-effective (World Health Report 2002, Jha
et al. 2006a, Asaria et al. 2007).
However, with respect to the decision to increase tobacco taxes, political con-
siderations have to be taken into account. Such considerations include, but they
are not limited to, concerns about the expected impact of a tax increase on: tax
evasion (smuggling) and tax avoidance; employment; inflation; affordability of
cigarettes and other tobacco products, especially for low income smokers: and the
relative prices of foreign and domestic brands. Furthermore, in some countries,
a culture of negotiated tax increases has developed between some governments
and manufacturers. Manufacturers responses to tax increases affect governments
expected revenues. Crucial to the success of the tobacco tax policy is an under-
standing of the political and economic environment in each country.
76 Chapter IV

4.1 Tobacco taxation and public health benefits

Growing evidence clearly shows that as taxes on tobacco products increase,

a significant number of premature deaths will be averted as youth are deterred
from taking up tobacco use and adult users quit, leading to substantial reductions
in the health and economic burden caused by tobacco use.
In India, for example, nearly one million people are expected to die prema-
turely from a disease caused by smoking by the early 2010s; these include deaths
from causes such as heart disease, cancer, respiratory diseases and tuberculosis.
Taxes on cigarettes are low in India, while taxes on bidis have historically been
close to zero. Significantly increasing these taxes would dramatically reduce the
prevalence of tobacco smoking and the death and disease it causes, while at the
same time raising substantial government revenues. Research shows that a 10%
increase in cigarette prices would reduce cigarette consumption by 3.4% in rural
India, while a 10% rise in bidi prices would reduce consumption by 9.2% and
8.5% in rural and urban India, respectively. These price increases would translate
to a 1.7% and 11.7% decrease in youth cigarette and bidi smoking prevalence,
respectively (John et al., 2010).
In terms of the health impact, a price increase of 52.8% on bidis through increased
taxes would avert about 4.6 million premature deaths among current bidi smokers,
while a cigarette price increase of 153% through increased taxes would avert an ad-
ditional 2 million premature deaths among current cigarette smokers. In addition,
by deterring the current cohort of Indian youth from initiating smoking, these
price increases would prevent an additional 1.6 million premature deaths caused
by cigarette smoking and 10.9 million premature deaths caused by bidi smoking.
In Russia, the tax increase based on the prospective tobacco excise law could
avert up to 80,000 deaths (about 0.4 percent of the expected tobacco-related
mortality in this cohort). However the number of smokers would be reduced
only marginally. If Russia chooses to raise tobacco taxes so that they account for
70 percent of the retail price, up to 2.7 million tobacco-related deaths among the
current Russian population could be avoided. This would reduce tobacco-related
mortality up to 12 percent with an even greater impact possible in the long run.
At the same time, the government would collect an additional RUB 153 billion
(US$6 billion) in excise tax revenue per year. (Ross et al, 2008).
In Ukraine, a relatively small tax increase that raises the tax to 50 percent
of the retail price could reduce the number of smokers by up to 500,000, avert
253,000 deaths (about 3.1 percent of the expected tobacco-related mortality in
this cohort), and annually generate about UAH 1.4 billion (US$ 281 million) in
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 77

additional excise revenues. If Ukraine were to raise tobacco taxes to 70% of the re-
tail price, the number of smokers would decline by almost two million, and about
one million tobacco-related deaths would be avoided in this cohort, reducing
tobacco-related mortality by 12 percent. At the same time, the government would
collect an additional UAH 4.2 billion (US$ 860 million) in excise tax revenue each
year. Taxes in Ukraine are low compared to neighbouring countries, creating an
incentive for smuggling duty-paid cigarettes out of the country. Therefore, a tax
increase in Ukraine would reduce incentives for illicit cigarette trade and reduce
duty-paid sales. However, even if all illegal cigarette exports are eliminated, tax
revenue would still increase by UAH 2.6 billion to 3.6 billion (US$ 539 million to
US$ 727 million), an increase of about 150 to 200 percent (Ross et al., 2009).
One has to recognize the highly political nature of tobacco control in general,
and tobacco taxation in particular, as well as the complex vested interests con-
cerned. Although the exact nature and extent of each actor and their interests
may be unique in each country, there are some widely used arguments used to
oppose tax increases. These include concerns about the effect of tax increases on
tax avoidance activities, smuggling, inflation, employment, poverty and protec-
tion of national industry. We turn to these issues now.

4.2 Tax avoidance and tax evasion

One of the challenges tax administrators face is how to sustain the revenue base
and flow, especially after a tax increase. The level of expected tax revenues de-
pends on limiting opportunities for tax avoidance and tax evasion, trends in
consumption, adoption of other tobacco control policies, and industry responses
to tax increases. Tax avoidance and tax evasion can make tobacco products more
affordable and more widely available and accessible, especially for youth and low
income smokers. Such activities undermine the health impact of higher tobacco
taxes and other tobacco control efforts.
Given the structure of the excise tax system and enforcement process, taxpayers
are faced with opportunities to reduce their tax payments. Any changes in the
tax system will induce different behavioural responses. For example, an increase
in tobacco excises may create an incentive to engage in tax avoidance and tax
evasion activities by both manufacturers and individuals, depending on enabling
environments (e.g. weak law enforcement and long judicial procedures, corruption
and weak governance) while encouraging some smokers to reduce consumption
(or discouraging others to take it up).
78 Chapter IV

Tax evasion should be distinguished from tax avoidance; tax avoidance is

legal, it is a change in economic or other activity, possibly at some cost, in order
to reduce tax payments. Tax evasion, however, involves illegal activities to avoid
tax payments.
There is a private cost to taking advantage of opportunities that reduce tax
payments. This cost may take the form of a change in consumption or purchase
behavior, an increasing probability of detection and penalty for evasion, and the
real resource costs of effecting avoidance and/or concealing evasion. These costs
depend on government policies that can be costly to implement, such as adminis-
tration and enforcement policies, but also on the setting of tax rates and tax bases.

4.2.1 Tax avoidance

Tax avoidance by consumers involves legal activities such as purchases for personal
consumption from a lower-tax jurisdictions or duty-free shops. For example,
smokers living in high tax jurisdictions may legally engage in cross-border shop-
ping in neighbouring low-tax jurisdictions, as happens in the US, the EU, and
other countries with significant population near borders (e.g. CIS countries, and
in Latin America, especially between Brazil and Paraguay (Ramos, 2009)). In some
countries, people may also buy cigarettes directly from other types of vendors
such as native reservations where some taxes are not applied.
The extent of cross border shopping and/or other tax avoidance activities by
individuals can be significant in some countries for example, in Luxembourg,
because of its low taxes and its proximity to large populations in higher tax coun-
tries. In practice, however, it is unlikely that individuals will travel long distances
at high cost just to buy cigarettes and save a modest amount of money.
The sale of duty-free tobacco products makes cheaper tobacco products more
readily available for consumption. This defeats the health purpose of taxation and
harms public health by encouraging personal consumption. The WHO FCTC
calls for a ban (or restriction) on the sale and import by international travellers
of tax and duty-free tobacco products, in order to increase the effectiveness of
tobacco taxation in reducing consumption. Eliminating duty free sales of tobacco
products will reduce opportunities for tax avoidance. There is growing evidence
of government and international actions to ban duty free sales (FCA, 2009).
Duty free tobacco product sales have been banned since 1999 to individuals
travelling within the EU; banned altogether in Romania (2010); and banned by
Bulgaria at land borders with non-EU countries. They were also recently banned
altogether in Nepal (2008). Since 2001 Canada has imposed a federal tax on
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 79

tobacco products sold in duty free stores (Canadian Cancer Society, 2010).
Just as duty-free tobacco product sales encourage consumption, so do allow-
ances for arriving travellers to bring in tobacco products duty-free and/or tax-free.
Although many countries still have a duty-free import allowance of 200 cigarettes
(or similar amount for other products), and sometimes even higher, an increas-
ing number of governments are eliminating or reducing the duty-free allowance
for arriving travellers. For example, in February 2010, the Hong Kong Special
Administrative Region Government announced, as a means to further protect
public health, a reduction in the limit to 19 cigarettes, meaning duty would be
required for an unopened package of 20 cigarettes (Hong Kong SAR Government,
2010). Countries like Barbados, Singapore and Sri Lanka do not permit any duty-
free allowances for cigarettes. In some EU countries duty-free import allowance is
restricted to 40 cigarettes (Bulgaria, Greece, Hungary, Lithuania, Poland Romania
and Slovakia).25 The amount is restricted to 80 cigarettes in Guatemala (Canadian
Cancer Society, 2010; European Commission 2009.)
Tax avoidance by manufacturers is less explored in the literature although
it does take place worldwide. It involves legal activities such as changing the
characteristic of the product, the package, the size of the production plan and
the pricing policy. For example, under specific taxation, manufacturers can ma-
nipulate the length of the cigarette or the size of the pack to reduce tax payment.
In some developing countries where multi-tiered tax systems are in place, we
observe various industry responses. In countries where the tier classification is
based on price level, for example, Egypt, Pakistan, Philippines, we observe that
prices of the brands tend to cluster near the top of each tier. To avoid a higher
tax, producers choose a different pricing policy to avoid a tax higher than the one
they might face in the presence of a single tax rate.
Some countries apply excise rates that vary with the type of the product
and/or the level of production. For example, in Indonesia the tax rates vary by
both the type of the product and the level of production. As lower rates apply for
lower levels of production, manufacturers can avoid higher taxes by establish-
ing a few smaller companies instead of a large production plant. Tiered tax rates
by production scale allow firms to avoid paying the highest tax, increasing profit
margins while reducing selling prices. When the tax rate depends on the type
of product, manufacturers may re-classify their product so that they are taxed
at a lower rate as seen recently in the United States where roll-your-own taxes

25 Applies in those countries (except for Romania) only for arrivals by land or sea (but not air)
from non-EU countries (duty-free sales within the EU are banned
80 Chapter IV

increased significantly compared to pipe tobacco taxes, leading to the repackaging

of roll-your-own tobacco as pipe tobacco. In general, under differential taxation,
there may be many ways to avoid tax. To eliminate tax avoidance, achieving higher
revenues and a larger health impact in the process, governments need to close
such loopholes in the tax law.
The degree and form of tax avoidance is of concern for several reasons. It con-
strains governments ability to raise revenue and control consumption through
taxation. Tax avoidance affects estimates of the level of smoking and price respon-
siveness when the analysis is based on sales data that are collected from country
cigarette tax receipts. As a tax rate increases, both taxed consumption and avoid-
ance activities change. Any estimate of the effect of tax on consumption will be
overstated if it fails to account for the triggered change in avoidance activities.
Governments need to prevent tax avoidance or at least control it. To do this, they
must frame tax rules so as to minimize opportunities for avoidance. In practice,
as governments amend legislation to close loopholes, tax advisers look for new
loopholes in the amended rules. Such loopholes are more likely to arise when
the tax structure is overly complex, as is the case in many developing countries.
Simplifying the tax structure will help reduce opportunities for tax avoidance
as well as monitoring costs per unit of revenue raised.

4.2.2 Tax evasion

Tax evasion usually involves taxpayers deliberately misrepresenting or concealing
their true economic activities to the tax authorities in order to reduce their tax
liability. For example, importers may evade customs duties and manufacturers
may evade domestic consumption taxes by under-invoicing or mis-declaration of
the quantity or description of the product. When the duty is ad valorem, under-
invoicing will reduce the tax base; when the duty is specific, mis-declaration
of quantity is more relevant.
Tax evasion, or illicit trade, involves both smuggling and illicit production. It may
involve genuine products or counterfeit. Smuggling is the trade of products
through unauthorized routes. It implies total or substantial evasion of customs
duties and excises, as well as income taxes. It can be long-distance, large-scale
organized smuggling or cross-border smuggling. Large-scale smuggling occurs
when large quantities of tobacco products are illegally transported, distributed
and sold without paying any tax at all, even in the country of origin. During
transport, export goods have in-transit status in which the goods can leave the
country of export without being assessed any taxes or duties. In-transit goods are
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 81

often temporarily stored in a country other than their final destination as they
await onward transfer. Large-scale smugglers often divert cargo at this point.
What gives rise to long-distance smuggling are the huge value differences between
export prices of major cigarette producing countries and the retail price of legal
cigarettes. Because taxes on cigarettes account for a large share of their price
relatively to other products (70-80% in the EU, 50-66% in some low and middle
income countries) and because tobacco products are relatively light, they are
especially appealing to smugglers.
Smugglers and legal traders may not always be two distinct groups. Smugglers
could be distributors camouflaging their smuggling with legal imports and reduc-
ing the costs of their legal imports with contraband (Fausti, 1999; Thursby and
Thursby, 2000). Major tobacco multinationals have been the subject of several
legal cases worldwide to determine the extent of their involvement: they were
accused of supplying the smuggled cigarettes or at least being aware of their
illegal destination.26
There is some evidence that the availability of duty-free sales of tobacco prod-
ucts has facilitated illicit trade in tobacco products in many countries. The evi-
dence includes government statements, internal tobacco industry documents
(an admission from British American Tobacco) and other reports on the issue.
(British American Tobacco, 2009; Collin et al., 2004; WHO 2009a; Canadian
Cancer Society, 2010). Cigarettes marked for duty-free sales may end up as con-
traband, often diverted into illegal distribution channels prior to even reaching
duty-free stores.
Reports from customs officials in countries have outlined the link between
duty-free and illicit trade. For example, according to the Organized Crime and
Corruption Reporting Project (OCCRP) (2008) a watchdog on organized crime
and corruption in Eastern Europe and Eurasia in July 2008, police officials in
Romania stated that half of all cigarettes smuggled into the country pass through
duty-free shops on the border. The Center for the Study of Democracy an in-
terdisciplinary public policy institute dedicated to the values of democracy and
market economy published in 2007 a short paper recognizing the link between
duty-free shops and increased smuggling in cigarettes in Bulgaria. Bulgaria (except
at the airport) and Romania have since banned duty-free tobacco product sales,
and reduced the duty-free import limit (for travellers from non-EU countries)
from 200 to 40 cigarettes in order to combat illicit trade (Sofia News Agency,
2010; Mediafax, 2010).

26 See, for example, http://www.ash.org.uk/smuggling/ or http://www.public-i.org/

82 Chapter IV

Several approaches have been used to obtain estimates of the extent of tobacco
smuggling, including relying on expert opinion, monitoring tobacco trade, com-
paring tobacco sales with total consumption estimated from survey data and
econometric modeling of the determinants of aggregate sales data (Merriman et al,
2000). Joossens et al. (2009) review a variety of estimates and conclude that 11.6%
of global cigarette market was illicit in or around 2007. A KPMG study, commis-
sioned by the European Commission, estimated that in 2004 illicit trade repre-
sented approximately 8-9% of the EU-25 tobacco market (Joossens et al., 2009).
With regards to econometric studies, there is no existing work on cigarette
large-scale smuggling in Europe and only one of bootlegging27 (Merriman et al.,
2000). Most of the evidence comes from North America (Baltagi and Levin, 1986,
1992; Thursby and Thursby, 1991; Galbraith and Kaiserman, 1997). Yurekli and
Zhang (2000) reveal significant long distance smuggling in the cigarette market
and its importance as a source of revenue lost. Worldwide, it is estimated that in
1995 approximately 6% of total tobacco products sold were smuggled through
diversion of untaxed exports from legal to illegal channels (Merriman et al., 2000).
Yurekli and Sayginsoy (2010) estimate that 3.4% of global cigarette consumption
in 1999 was smuggled.
To evaluate the size of the informal tobacco sector, let alone its composition,
is difficult, especially as it evolves over time. In 2000/01 in the UK, most illicit
cigarettes were genuine, locally manufactured products, exported to continental
Europe and then smuggled back to the UK. In 2002 and 2003, leading UK tobacco
manufacturers signed the Memoranda of Understanding under which they agreed
to control the supply chain. These agreements were voluntary and non-binding,
and as such their effectiveness depended on the manufacturers goodwill. In 2006,
the UK introduced changes in its legislation, setting high penalty payments. As a
result of these measures, smuggling of UK genuine brands was reduced. However,
this type of smuggling was replaced by smuggling of counterfeit and cheap non-
UK brands. Looking at other tobacco products, smuggling in hand rolling tobacco
(HRT) remained a serious problem: more than half of HRT consumed in the UK
is illegal (ASH, 2009). There is still scope for improving the supply chain control.
Illicit production may involve production of genuine brands by legal manu-
facturers who declare only a fraction of their production to the tax authorities.
This form of tax evasion is prevalent among large cigarette producing coun-
tries such as Egypt, India (Bidis), Indonesia, Russia, Pakistan and Philippines.

27 Bootlegging involves the purchase, by individuals or small groups, of tobacco products

in low tax jurisdictions, in amounts that exceed customs limits, for resale untaxed in high tax
jurisdictions (Joossens et al., 2009).
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 83

It may involve production of counterfeit products by illegal domestic manufac-

turers. This occurs in, for example, Russia and South-East Asia, with most of the
counterfeit cigarettes coming from China. In 2007, three reports concerning the
discovery of illegal plants for cigarette production in Austria, the Czech Repub-
lic, and Slovakia were submitted to the World Customs Organization (WCO,
2007). Strengthening cooperation, exchange of necessary information. and grant-
ing greater investigative powers to Customs services may result in dismantling
of more illegal manufacturing lines.
It is usually the size and composition of seizures that give us an idea of the
composition of the illicit market. However, seizures may not be representative
of the illicit market as a whole. Moreover, making comparisons across countries
on the basis of seizures is not meaningful, as, for example, customs investigative
techniques, reporting procedures and law enforcement differ.
The presence of an illicit market, especially if it is of a considerable size, has an
impact on both consumption and tax revenues. If smuggled cigarettes account
for a high fraction of the total market, the average price of all cigarettes will fall,
leading to an increase in consumption. As illicit tobacco products become more
available, their share in individual consumption will increase and the average
price paid by smokers will decrease. Apart from affecting consumption by cur-
rent smokers, the price decrease affects potential future smokers, as individuals
are more likely to take up smoking the lower the price. Evidence shows that
those who buy illicit tobacco products are more likely to be young and belong to
semi-skilled and unskilled occupation groups, as these groups are found to be
more price sensitive (West, 2008). As a result, higher consumption will contribute
to higher mortality from smoking-related diseases.
High tax increases may provide financial incentives for smuggling, especially
when enforcement and tax laws are weak, penalties are small, and it takes a long
time to prosecute smugglers. Literature does not provide clear cut results on the
effect of commodity tax increase on total sales and tax evasion, in noncompeti-
tive environments (e.g. Thursby et al, 1991; Thursby and Thursby, 2000) or on
the relative effects of specific and ad valorem taxes (Delipalla, 2009a, 2009b).
It is clear, however, that an increase in penalties or detection probability has a
clear negative effect on tax evasion. In practice, corruption often renders control
of evasion difficult. Moreover, as corruption reduces the expected cost of smug-
gling, it encourages it. Some governments have resorted to privatization of tax
enforcement to enhance efficiency of the tax system, the assumption being that
leakage of revenue will be smaller under a privatized regime. In Bangladesh, for
example, a part of Customs administration was privatized as early as 1991.
84 Chapter IV

Governments should require identifying information to be included on all

tobacco products produced domestically so as to facilitate tracking and tracing
of these products through the distribution process and should work with others
in the region to adopt similar requirements. This information would be highly
useful in enforcement efforts, and allow Customs to identify illicit products more
easily and to identify those higher up in the distribution chain that are responsible.
Severe administrative penalties should be imposed on those caught engaging
in illicit trade so as to significantly increase the swiftness and severity of these
penalties, making them a greater deterrent.
Moreover, measures of the extent of illicit tobacco product availability and
pricing should be incorporated into a broader industry surveillance system in
each country. Reliable measures would reduce Customs authorities reliance on
the tobacco industry for estimates of the extent of illegal trade in their country.
Spain provides a good example of effective measures to control the supply of
smuggled tobacco. Investments in strengthening intelligence, increasing customs
activity in border areas, and developing international collaborations targeting
smuggling rose from 4 million in 1993/94 to almost 40 million in the period
1996-2000 (Joossens and Raw, 2008). As a result, the market share of smuggled
cigarettes fell from 16% to 2%, and tax revenues increased from 2300 million
to 5200 million, equivalent to 68 in tax revenue for every 1 spent on anti-
smuggling measures (ASH, 2009).
In 2000, the European Commission (EC) took a number of tobacco companies
to court accusing them, among other things, of smuggling. In 2001, ten European
countries led by Italy joined the lawsuit. In 2004, the case against Phillip Morris
International (PMI) was dropped as PMI agreed to pay the EC $1 billion over
12 years and to control future smuggling of its brands. PMI developed a special
tracking and tracing system and marked 200 million master cases with unique
codes. Italys illicit trade in cigarettes fell from 15% in the 1990s to 1-2% in 2006
(Joossens and Raw, 2008). Since 2008, PMI introduced tracking and tracing at the
carton level in Eastern Europe. Japan Tobacco International (JTI) signed a similar
agreement in 2007. In 2009, the UK joined in signing anti-smuggling agreements.
Recognizing the importance of strong international cooperation to eliminate
illicit trade in tobacco products, the Parties to the WHO Framework Conven-
tion on Tobacco Control (WHO FCTC) created a negotiating body to develop
a protocol on illicit trade in tobacco products. Negotiations started in February
2008 and are ongoing. A draft of the text of the protocol will be presented at the
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 85

fourth session of Conference of the Parties to the WHO FCTC in November 2010
for their consideration. The current draft of the protocol includes provisions to
control the tobacco supply chain, measures to define offences and set sanctions,
measures to facilitate international cooperation and data sharing and institutional
measures with regards to the Protocol itself. The main elements of the tobacco
supply chain section are:28
Licensing (required for all engaged in manufacturing of tobacco products but
also in manufacturing equipment, commercial activities, transportation and
primary processing of tobacco products)
Customer identification and verification (due diligence)
Tracking and tracing (affixing secure and non-removable markings on tobacco
products and manufacturing equipment used in the manufacturing of local
and imported tobacco products)
Record-keeping (of activities of those engaged in the commercial sale of tobacco
or in the manufacture, sale, distribution, storage, shipment, import or export
of tobacco products or manufacturing equipment used in the manufacture
of tobacco products)
Security and preventive measures (to ensure compliance with regulation)
Banning or ensuring compliance to obligations of the Protocol in the internet
and other telecommunication-based modes of sale
Limiting, licensing or prohibiting tobacco in free-trade areas and for duty-free
sales (major sources of illicit tobacco trade).29
Although all forms of tax avoidance and tax evasion may affect revenues and
tobacco control, policy makers need to know their absolute and relative impor-
tance when deciding whether and how to allocate resources to prevent them.
For example, when both border crossing and large scale smuggling is present,
border crossing might be considered less harmful than smuggling because,
although it encourages consumption, causes unnecessary transportation costs, and
shifts tax revenues between governments, it is legal if the quantities purchased fall
below specified limits. Smuggling, in contrast, is illegal and, apart from encourag-
ing smoking, it may direct revenue to criminal organizations and generate costs
associated with violence or law enforcement.

28 Source: http://www.who.int/fctc/inb/en/
29 A recent study demonstrates that the benefits from implementing the protocol in the UK
are highly likely to exceed the costs (ASH, 2009).
86 Chapter IV

4.3 Protecting Domestic Brands

Until the mid-1990s, governments in many countries were the sole producers of
a variety of products including tobacco products. One of the main reasons for
governments involvement was to provide affordable products for mass popula-
tion. Today, with the exception of a few countries, government owned tobacco
industries have been privatized. China, Thailand, Egypt (52% still owned by the
government), Viet Nam, Japan (less than 49%), Moldova, and Iran still maintain
full or partial control of tobacco manufacturing and distribution. Historically,
cigarettes produced by government owned companies have been priced much
lower and used lower grades of tobacco than foreign brands.
Currently, governments that impose a differential excise system often levy
higher taxes on premium or high price brands, often produced by foreign manu-
facturers, than they do on lower grade, lower priced brands that are often produced
domestically. As taxes increase, premium and high-price brands are expected to
generate more stable revenue than the other price bands due to their less price
sensitive consumption base. High income smokers are more likely to smoke pre-
mium, high price brands and are less responsive to price than are smokers in
lower income groups. Given their market share and the high taxes that are ap-
plied to them, premium brands generate a relatively high share of total tobacco
tax revenues in various countries, as shown in Table 7.

Table 7. Excise revenue by price band, share in tobacco excise revenues and sales, 2008.

PAKISTAN EGYPT TURKEY

Excise Share* Share GST Share* Share Excise Share* Share
Mil. in in Sales Mil. LE in GST in Sales Mil.TL in in Sales
Rs** Excise % ** Rev. % % Excise %
Rev. % Rev %
Premium 11,231 29 10 832 12.2 6.6 3,129 28 20
Mid price 24,266 63 79 990 14.6 14.5 4,396 40 40
Economy 2,744 7 10 4,983 73.2 78.9 3,591 32 41
Total 100 100 100 100 100 100

* Share in tobacco excise revenue.

** Excise revenue includes Federal and State excise duties in Pakistan. Excise revenue is General
Sales Tax in Egypt
Sources: Authors calculations using data from MoF Egypt (2009), FBR Pakistan (2009) and Yurekli
et al. (2010)
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 87

Governments also have a tendency to keep the prices of tobacco products

consumed by the majority of population relatively lower, by either not taxing
these products or by keeping the tax rates on these products significantly lower.
This is especially the case for bidis and smokeless tobacco in India (Sunley, 2008;
Goodchild, forthcoming), papirosy and non-filtered cigarettes in Russia (Ross
et al,, 2008), and waterpipes in Egypt (MoF Egypt, 2009). In some cases, due to
low consumption level, governments impose either no or very low tax on some
products (e.g. loose tobacco). Consequently, as the tax gap increases, consumers
switch towards those products, as is the case for example in Viet Nam (Guindon
et al., 2010) and Poland (WHO, 2009b).

4.4 Tobacco taxes and affordability

To the extent that governments decide to use higher tobacco taxes to reduce the
health and economic consequences of tobacco use, they need to consider more
than just the absolute level of taxes. Changes in the prices of other goods and
services need to be taken into account. Increases in taxes on tobacco products
that do not result in increases in prices that are larger than the increase in other
prices will result in a drop in the prices of tobacco products relative to other
goods and services (a drop in the real or inflation adjusted price). Rising nominal
but falling real prices for tobacco products will lead to increases, not decreases,
in tobacco use and its consequences.
The U.S. in the 1970s provides a clear example of this. Despite continually
increasing and well disseminated information about the health consequences of
smoking, a new health warning label on cigarette packaging and advertising, a
ban on broadcast advertising for cigarettes, the spread of restrictions on smoking
in public places, including restaurants and workplaces, and an increase of over
53 percent in nominal cigarette prices, per capita cigarette consumption rose by
11.4 percent from 1970 to 1979. The increased consumption was caused by a 16
percent fall in the real prices of cigarettes during this period, largely the result of
no increase in the countrys specific tax at the national level and small increases
in specific taxes in some states that were not enough to keep pace with inflation.
Some countries that use tobacco taxes as a way to reduce tobacco use and
improve public health have addressed this problem by adopting policies that
automatically increase their specific tobacco taxes so as to keep up with inflation
and maintain their real value. Australia, for example, adjusts its cigarette taxes
twice each year so that the inflation adjusted value is maintained.
88 Chapter IV

Similarly, the impact of income on tobacco use needs to be considered when

evaluating the affordability of tobacco products. In most countries, particularly
low- and middle-income countries, consumption of tobacco products increases
as incomes increase. As a result, the reductions in tobacco use caused by tobacco
tax increases may be more than offset by the increases in tobacco use that result
from higher incomes. While this would result in a larger increase in tax revenues
than would result from the increased tax alone, it also implies an increase rather
than a reduction in tobacco use and its consequences.
This illustrates the importance of reducing the affordability of tobacco products
when a key goal of tobacco taxation is to reduce tobacco use, given that afford-
ability depends on both price and income. As Blecher and van Walbeek (2004;
2009) show, in high income countries tax and price increases have generally
outpaced growth in incomes, so that the affordability of cigarettes has, on aver-
age, declined considerably since 1990, contributing to the reductions in smoking
that have occurred in these countries. In contrast, affordability of cigarettes (and
almost certainly all other tobacco products) has increased significantly in low and
lower middle income countries where tax and price increases have been modest
and well below increases in incomes. Figure 12 shows cigarette affordability over
time in 5 countries. Using 1995 as the base year, estimated values greater (less)
than 1 indicate that cigarettes are less (more) affordable relative to 1995.

Looking at China in particular, Hu et al. (2008) show that, despite a more

than doubling of real cigarette prices between 1990 and 2005, cigarettes became
more than twice as affordable because of the sharp growth in income in China
during this period. Consistent with economic theory, one result of this increased
affordability is that the demand for cigarettes in China has become much more
inelastic (less sensitive to price changes) over time. Moreover, the increased af-
fordability of cigarettes led to about a nine percent increase in per capita cigarette
consumption in China during this period. To date, no country has adopted a
policy that automatically adjusts tobacco product taxes in order to prevent them
from becoming more affordable over time as incomes increase.
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 89

Figure 12. Cigarette affordability in five countries.

Less Affordable
than 1995

Affordability
Index
1995 = 1

More Affordable
than 1995
1990
1991
1992
1993

1994
1995

1996
1997
1998
1999
2000
2001
2002

2003
2004

2005
2006
2007
Bangladesh China Indonesia Pakistan VietNam

Note: The affordability index is the ratio of the price of the most popular brand to per capita income.
Sources: WHO GTCR 2009

4.5 Tobacco taxes and tobacco product substitution

Tobacco tax rates that differ across products, and tobacco tax changes that affect
prices across products differently, will lead to some substitution among these
products (Chaloupka et al., 2000). For example, in Poland cigarette tax increases
leading up to the countrys accession to the European Union led some smokers
to switch from manufactured cigarettes to roll-your-own tobacco (RYO). This
led to subsequent increases in the RYO tobacco tax to bring it closer to the tax
on manufactured cigarettes, along with further increases in both taxes. However,
other tobacco product taxes increased modestly by comparison, leading to further
substitution this time to pipe tobacco, which many consumers used to make
cigarettes rather than smoking it in pipes. The most recent Polish tobacco tax
increases (in March 2009) addressed this by bringing the pipe tobacco tax up
to the same level as the RYO tax.
90 Chapter IV

Taxing tobacco products consistently so that the tax accounts for a comparable
share of price on different products and so that tax increases result in proportion-
ate increases in the prices on all products reduces the potential for substitution
among these products. However, one has to take into account the extent to which
the price elasticity of demand varies among different tobacco products, which
products are close substitutes (cross price elasticities), as well as the starting tax
rates on each tobacco product.
As the starting tax rates may be very low (or even zero) for some tobacco prod-
ucts, substantial tax increases to reach a tax share in price that is comparable with
other tobacco products may prove to be difficult to implement politically. In India,
bidis are consumed by relatively poorer individuals. As a result, bidi consumers are
much more sensitive to price changes (e.g. exhibit a much higher price elasticity
of demand) compared to cigarette smokers. Different price elasticities of demand
among tobacco products mean that the same proportionate change in price across
these products will lead to different changes in consumption.

4.6 Tobacco taxes and poverty

Concerns about the burden of tax increases on the poor are another barrier
to higher tobacco taxes. Indeed, in some countries, tobacco tax levels and struc-
ture are in part designed to produce low prices on some brands or products
in order to keep them affordable for poor users. Rather than being pro-poor,
a policy like this results in greater tobacco use among those on lower incomes.
As a consequence, the poor end up bearing a disproportionate share of the health
and economic burden of tobacco, with differences in tobacco use among the rich
and poor accounting for much of observed socioeconomic differences in health
(Bobak et al., 2000). Moreover, tobacco use can increase poverty as funds are
diverted from spending on basic necessities like food, housing, education and
health care to spending on tobacco products (Nargis et al., forthcoming). Figure
13 shows that the health consequences from smoking are much higher among
lower socio-economic group in selected countries, leading to higher death in
these groups and accounting for much of the health gap between the rich and the
poor. This is exacerbated by family income losses that result from missed work
time due to diseases and premature death caused by tobacco use and increased
spending on health care to treat illnesses caused by tobacco.
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 91

Figure 13. Differential health outcome due to smoking.

Risk of dying
at ages
35-69 years
England USA Canada Poland
25% and Wales 22%
20% 19%

14% 15%
15% 13%
10% 10%
10% 8%
6% 5%
5% 4% 4%

0%
Social Class Education Neighbourhood Education
income
High (I / II)

Mid (III / IV)

Low (IV)

High
(>12 years)
Mid
(12 years)
Low
(<12 years)

High (20%)

Mid (60%)

Low (20%)

High
(>12 years)
Mid
(12 years)
High
(<12 years)
Note: Social inequalities in male mortality in 1996 from smoking. Values are percentages of
35-year-old men dying at ages 3569 years from smoking if the population death rates of 1996
were to remain unchanged.
Source: Jha et al., 2006b

Whether or not tobacco taxes fall more heavily on the poor depends on several
factors, including tax structure and tobacco use patterns for those at different
income levels. Tobacco taxes will generally be regressive when prevalence of to-
bacco use and consumption patterns are similar across income levels and when
taxes are similar across tobacco products, given that tobacco taxes paid will ac-
count for a greater share of income for the poor than for the rich. The regressivity
of tobacco taxes will be more pronounced in countries where tobacco product
consumption is greater among the poor than among those on higher incomes.
However, tobacco taxes can be less regressive or even progressive in countries
where consumption levels increase with income and/or where higher taxes are
applied on the products consumed by higher income consumers.
Similarly, whether or not tax increases will fall more heavily on the poor de-
pends on how tobacco use among the poor and rich changes in response to the
tax increases. Consistent with economic theory, studies from a growing number
of countries generally find that there are considerable differences in price elastic-
ity of tobacco use among socioeconomic groups in a given country, with tobacco
use in lower income populations much more sensitive to price than tobacco
92 Chapter IV

use in higher income populations. For example, Sayginsoy et al. (2002) estimate
cigarette demand elasticities of -1.33, -1.00 and -0.52 for low, middle and high
income populations in Bulgaria. Similarly, van Walbeek (2002) estimates elastici-
ties by income quartile ranging from -1.39 for the lowest quartile to -0.81 for the
highest quartile in South Africa. In Indonesia, Adoietomo et al. (2005) estimate
cigarette demand elasticities of -0.67, -0.33 and -0.31 for low, middle and high
income populations. These estimates imply that a tax increase will reduce tobacco
use most among the lowest income populations while having less of an impact
on higher income populations.
As lower socio-economic groups have lower response to health education than
higher socio-economic groups, increases in the real cost of cigarettes, through
taxes, will help reduce differences between different socio-economic groups in
prevalence of smoking and smoking-related diseases (e.g. Townsend et al, 1994).
Given these findings, even if the tobacco tax itself is regressive, a tobacco tax
increase can be progressive. Based on existing evidence, Nargis and colleagues
(forthcoming) summarize this for Thailand, Bulgaria, and Turkey. They show that
because of differences in price responsiveness across income groups, increases
in cigarette taxes lead to a reduction in the overall share of tobacco taxes paid by
the lowest income groups in each country, while the share paid by the highest
income groups increases. Moreover, because of the relatively larger reductions
in tobacco use among the poor, they will gain more of the health and economic
benefits that result from the tax increase.
Moreover, when one accounts for self control problems that individuals do
not make optimal tradeoffs between the immediate gratification they get from
consumption now and their long run desires that result in overconsumption
of tobacco products, and accounts for the benefits from reduced consumption,
taxes that appear regressive are less so and may even be progressive (Gruber
and Koszegi, 2008). This is more likely as there are greater differences between
the poor and rich in the responsiveness of tobacco use to price; as the poor are
more responsive, the benefits that accrue to them from tax-induced reductions
in consumption will be larger than those that go to the rich.
Gruber and Koszegi (2008) demonstrate this for the U.S., where those in the
poorest income quartile spend ten times as much of their incomes on cigarettes
as do those in the top income quartile, and where they estimate that cigarette de-
mand among the poor is much more responsive to price than demand among the
rich. In this case, for plausible assumptions about the extent of time inconsistency
in smokers behavior (the extent of the difference between the taste for immedi-
ate gratification and long run preferences), cigarette taxes are quite progressive.
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 93

Given that differences in spending on tobacco products by income are less pro-
nounced in most low and middle income countries, and given the evidence from
these countries that demand among the poor is more sensitive to price than de-
mand among the rich, tobacco taxes are likely to be even more progressive.
Finally, to the extent that there are continuing concerns about the impact of
tobacco tax increases on the poor, governments can address these concerns by
using the new revenues from a tax increase in a way that provides greater benefits
to the poor. In this sense tobacco taxation becomes a pro-poor policy. A grow-
ing number of governments do this by dedicating some portion of tobacco tax
revenues to programmes targeting the poor. For example, Egypt is considering
increasing taxes on cigarettes and use the revenue generated to widen the coverage
of health insurance and improve health services among the poor. Also, following
the recent tax increase in Turkey, the government is considering using a portion
of the extra revenues to increase health coverage and improve health services,
which will benefit the poor.

4.7 Tobacco tax increases and inflation

At times the inflationary impact of cigarette and other tobacco product tax in-
creases is raised as an argument for not increasing these taxes. This may be par-
ticularly true in countries where wages and/or a significant share of government
spending is indexed to inflation (e.g. for public pension payments) and/or where
government policy is to keep inflation low.
The extent to which tobacco product tax increases lead to increases in inflation
depends on several factors, most notably the share of these taxes in prices and
the weight tobacco prices are given in computing a price index. For example, if
taxes account for 25 percent of tobacco product prices, a doubling of the tax (100
percent increase) will increase prices by 25 percent. If the weight given to tobacco
products in the price index is three percent, the index will rise by 0.75 percent in
response to the tax increase. As tobacco taxes account for a larger share of tobacco
product prices, the inflationary impact of a tax increase will be greater. Similarly,
as tobacco products are given more weight in computing a price index, a given
tax increase will have a greater inflationary effect. In general, for most countries,
the inflationary impact of tobacco product tax increases will be relatively small.
The generally small impact of tobacco taxes on inflation is illustrated in Table
8 where various combinations of tax levels (as a percent of price) and tobacco
weights in the price index are examined.
94 Chapter IV

Table 8. Inflationary impact of tobacco tax increases.

Tax as a share of price Tobacco weight in price index Inflationary impact

Low Medium High Low Medium High Low Medium High
(<40%) (40-70%) (>70%) (<2%) (2-4%) (4-8%) (<1.0%) (1-2.5%) (>2.5%)
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X

Note: Midpoints of ranges for tax and tobacco weight are used for computing inflationary impact.
Source: Authors simulations

Consumer price indices have multiple purposes. They are an important eco-
nomic indicator for most countries and are often a key determinant of monetary
policy. Inflation rates directly impact on interest rates and exchange rates. In many
countries, changes in wages, social security benefits, and other payments are tied
to inflation, as measured by a price index. In some countries, various taxes are
linked to price indices; for example, US income tax brackets are adjusted annually
to reflect changes in consumer prices, while Australia and New Zealand regularly
increase their cigarette taxes to keep pace with inflation. Price indices are used
to provide more accurate comparisons of changes in expenditures, incomes and
prices for specific goods over time as well as to allow comparisons across countries.
Given the many uses of consumer price indices and the potential inflationary
impact of tobacco tax increases, some governments have developed alternatives
that exclude tobacco (and sometimes other goods) for some uses. For example,
since 1992, France has excluded tobacco products from the price index used for
adjusting minimum wages. Given its utility for indexing various payments, some
governments exclude prices for a variety of products they consider unnecessary
or inappropriate, including those for alcoholic beverages, gambling, and tobacco.
For example, since 1991, Luxembourg has excluded tobacco products, hard liquor,
and certain services closely linked to sliding wage scales from its consumer price
index. To date, however, while many countries do report consumer price indices
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 95

that exclude tobacco products, their most widely used indices including those
used for indexation of wages, pension payments, and other outlays continue
to include tobacco products.
To the extent that concerns about their impact on inflation are a barrier to
tobacco tax increases, excluding tobacco products from the basket of goods used
in developing key price indices would greatly reduce these concerns. In addition,
some have observed that the inclusion of tobacco products in key price indices
results in a distorted measure of price for many consumers, particularly in coun-
tries where a small and declining minority of the population use these products.
Likewise, given that the weights used to compute price indices in many countries
change infrequently, the inflationary impact of tobacco product tax increases will
be overstated as consumption of these products falls in response to tax increases.
Finally, some have suggested that excluding tobacco products from price indices
would increase the public health impact of tobacco tax increases by providing less
of a cushion for users whose wages or benefit payments are indexed (Alchin, 1995).

4.8 Tobacco taxes and employment

Opponents of tobacco tax increase often suggest that the tax increases will result
in job losses, noting that many are employed in tobacco growing, manufacturing
and distribution. However, as Warner (2000) has noted, an economic presence of
tobacco does not imply an economic dependence on tobacco. Many of the jobs
that are counted in estimates of the economic contribution of tobacco are far
from dependent on tobacco, but rather involve tobacco in some limited way, often
indirectly (e.g. retailers who sell tobacco products, among many other products,
or jobs in the heavy equipment sector where farming equipment is produced).
Similarly, these estimates include so-called expenditure induced employment
jobs that result from spending by those whose incomes are earned in the jobs
counted as tobacco related. In general, only jobs in tobacco farming (which are
often part time and for which tobacco is one of several crops), tobacco leaf drying
and warehousing (which involves very few jobs), and tobacco product manufac-
turing can be considered truly dependent on tobacco.
In most countries, employment in tobacco dependent sectors has been fall-
ing over time as farming techniques have improved and as tobacco product
manufacturers have adopted new, more capital intensive production methods.
In some countries, increased imports of tobacco leaf and/or tobacco products
have contributed to reduced domestic employment in tobacco dependent sectors.
96 Chapter IV

For most countries, the job losses in tobacco dependent sectors that have resulted
from these factors exceed any job losses resulting from higher taxes and other
tobacco control efforts. (Lei et al., forthcoming).
More importantly, any tobacco dependent jobs lost in response to the reduced
demand for tobacco products caused by higher tobacco taxes will be offset by new
jobs in other sectors. The money not spent by tobacco users who quit or spend less
on tobacco products after a tax increase will not disappear from the economy, but
will instead be spent on other goods and services, creating jobs in these sectors.
For example in India, the impact of higher taxes on employment is not expected
to be significant, given Indias growing economy and an expected slow reduction
of tobacco-related jobs concurrent with increases in jobs in other sectors as funds
once spent on tobacco are spent on other goods and services (John et al., 2010).
Similarly, government spending of the new tax revenues that result from a tax
increase will create jobs in other sectors. Study after study has demonstrated that
increases in tobacco taxes or implementation of other tobacco control measures
do not lead to net job losses; in many countries, such efforts result in net increases
in jobs as spending is shifted to more labour intensive goods and services (Lei, et
al., forthcoming; Jacobs, et al., 2000). This is particularly true for countries where
significant shares of tobacco leaf and/or tobacco products are imported, given
that much of the money spent on tobacco products will flow out of the country,
in contrast to the spending that replaces spending on tobacco in response to tax
increases or other tobacco control measures.
Even global tobacco tax increases are unlikely to have a significant impact on
tobacco dependent employment in most countries. For a few agrarian countries
that do depend heavily on tobacco leaf exports (e.g. Malawi), a sharp, immedi-
ate reduction in global demand for tobacco products would lead to significant
job losses in the short run. However, given the current upward trend in global
demand, higher taxes and other tobacco control measures are not likely to result
in a sharp drop in demand in the short run, but rather a slowing of the increase in
the near term followed by slowly falling demand in the longer term. This implies
that any job losses in these countries will not happen for many years, allowing
for a gradual transition from tobacco to other crops.
Countries that are concerned about the impact of tobacco tax increases on
domestic employment in tobacco dependent sectors can alleviate these concerns
by adopting programmes that would ease the transition from tobacco farming and
manufacturing to other economic activity. Crop diversification programmes that
support farmers and retraining programmes for those involved in tobacco product
manufacturing could easily be funded by a small portion of the new revenues that
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 97

result from increases in taxes on tobacco products. In Turkey, for example, the
government sponsored alternative crop programme that was implemented in
anticipation of the privatization of the countrys cigarette monopoly has proven
effective in moving many tobacco farmers to other crops (Yurekli et al., 2010).

4.9 Tobacco taxation and harm reduction

A wide variety of tobacco products are on the market today, with new prod-
ucts seeming to emerge continuously (see www.tobaccoproducts.org for more
details). These products can be grouped into two broad categories combustible
(smoked) products and non-combustible (usually used orally) products. In some
countries, a range of both products have been available for many years, and,
in a few, manufactured cigarettes account for a relatively small share of over-
all tobacco use. For example, in India, many more tobacco smokers use bidis
(dried tobacco hand-rolled in a tendu leaf) than manufactured cigarettes, while
a large portion of the population chews tobacco in the form of paan masala or
gutka. In Indonesia, kreteks (clove cigarettes) are widely smoked, while in many
Middle Eastern countries, waterpipe smoking of tobacco is common (e.g. hookah
or shisha smoking).
In recent years, the variety of available products has expanded considerably,
particularly in high-income countries, as the tobacco industry has introduced
products that are marketed as reduced risk products. Some new cigarettes, for
example, claim to reduce the carcinogens contained in their smoke while oth-
ers deliver considerably less tar, nicotine and/or carbon monoxide. Many new
non-combustible products are being similarly marketed, from Swedish Matchs
snus (a moist snuff product that uses tobacco cured in a way that is supposed
to significantly reduce cancer causing agents) to the lozenges, dissolvable strips,
tobacco chewing gum, and others. At the same time, the number of available
non-tobacco products that deliver nicotine has risen, ranging from those intended
for smoking cessation (nicotine gum, patches, inhalers, etc.) to the e-cigarette
(a battery powered device that delivers nicotine through a mixture of air and
water vapor).
Governments have struggled with how to regulate these products and, given
experiences with filtered and low-tar and nicotine cigarettes, have been reluctant
to allow these products to be marketed as less harmful. Research has clearly
demonstrated that smokers perceptions that low-tar and nicotine cigarettes, for
example, were safer than regular cigarettes led many who might have otherwise
98 Chapter IV

quit smoking to continue. Only decades after their introduction did it become
clear that the machine measurements of tar and nicotine did not reflect human
exposure and that these cigarettes were not safer than regular cigarettes.
The variety of tobacco products available have led some to suggest that tobacco
excises be set differentially, so as to more heavily tax those that have greater
health risks, while taxing those perceived to be safer at lower levels (or not at all).
Harris (1980), for example, suggested that a differential tax based on tar and
nicotine content could promote public health by encouraging smokers to move
from high tar/nicotine brands to low tar/nicotine brands, assuming that the latter
were less harmful. However, given what we now know about the relative risks
of these cigarettes, its clear that such a policy would have done more harm than
good as it would have likely kept even more smokers in the market consuming
what they perceived to be safer products.
To date, differential taxation of various tobacco products (e.g. for filtered
vs. unfiltered cigarettes or for smoked vs. smokeless products) does not seem mo-
tivated by interests in promoting harm reduction. Where differential taxes exist,
they appear more motivated by efforts to protect domestic producers (e.g. those
producing unfiltered cigarettes) from multinational firms (e.g. those producing fil-
tered cigarettes) or by efforts to increase revenues (e.g. by taxing the manufactured
cigarettes consumed by higher income, less price sensitive consumers more than
the hand-rolled bidis smoked by more price sensitive, lower income smokers).
Recognizing past misrepresentations and current uncertainties, at this point
in time, designing a tobacco tax system that favours products perceived to be
safer while disfavouring those perceived to be more harmful should await clear
evidence of a harm reduction benefit for both the individuals using the products
and the public health of the general population.

4.10 Tobacco tax revenues, health expenditure and earmarking

Financing the health-care system is crucial in most countries as it serves to im-

prove health care access and the quality of the services provided. This also reduces
the risks of high economic costs due to disease and consequent death. In low- and
middle-income countries, financing has become a central issue of health reform,
given the large proportion of out-of-pocket expenses on health and the financial
constraints this imposes on poor households (Prakongsai et al., 2008).
The use of government tax revenues to pay for health services is a fairly recent
innovation in health care financing. Until the mid-twentieth century, the major
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 99

alternatives to out-of-pocket payments for health care services were private phi-
lanthropies, mutual associations or social insurance plans (e.g. sickness funds)
(WHO, 2004). In the case of tobacco products, earmarking (through passing a law)
or dedicating (commitment by the Government but no legislation needed, which
is more flexible than earmarking) revenues from tobacco taxes for health purposes
can be seen as a way to correct for the negative health consequences of tobacco use.
Earmarking can be classified according to two criteria. First, according to
the link between the tax and the expenditure it finances: a strong or tight link
implies that all or most of the revenue goes towards financing a particular ex-
penditure, and that the expenditure does not benefit (significantly) from other
financing sources (e.g.the general fund). A weak or loose link implies that only
a portion of the proceeds of the tax finances the expenditure in question, and/
or the expenditure benefits (significantly) from other financing sources. Second,
according to the type of expenditure benefiting, earmarking can be specific/nar-
row (e.g.a service provided by a public enterprise), or broad/wide (e.g.social
security, education). The main argument against earmarking is that it may intro-
duce rigidities in the budgetary process that limit the use of funds for alternative
purposes, discouraging the optimal allocation of resources and hence reducing
social welfare.
Buchanan (1963), starting with the median voter-taxpayer as the decision
maker in the tax-spending process (instead of the fiscal authority), showed that
earmarking can be desirable. If voters are offered a series of public goods/services
with each financed by a corresponding tax, the outcome of their choice is likely to
reflect their preferences better than voting on a package of expenditures financed
by a general fund. Since Buchanans seminal work, a number of economists have
shown why certain types of earmarking can be desirable or indeed observed
in practice. For example, Pirttil (1998) argues that earmarking revenue from
a corrective environmental tax to compensate those who suffer the most from
such a tax may be desirable. Marsiliani and Renstrom (2000) show that earmark-
ing can act as a commitment mechanism where there is a time-inconsistency
problem in environmental tax policy: future politicians can be prevented from
eliminating the tax or reducing it because its use is earmarked for a desirable
expenditure programme. Along the same lines, Brett and Keen (2000) explain
earmarking as a means by which a weak incumbent politician locks in the use
of certain tax revenues (from environmental Pigovian taxes) and prevents fu-
ture politicians from altering that use. Dhillon and Perroni (2001) justify ear-
marking on the basis that it improves the monitoring of government spending
by private individuals.
100 Chapter IV

Earmarking in modern public finance finds its strongest support in the prin-
ciple of benefit taxation and user fees. According to this principle, tobacco
taxes must be paid by those who benefit from tobacco-related health services,
a condition that is impossible to satisfy as not all tobacco smokers suffer from
tobacco-related diseases, and tobacco tax revenue may not be enough to finance
spending needs. It could be argued, however, that the tax can take the form
of a compulsory health contribution to finance a health insurance programme
for tobacco-related diseases. There are two weaknesses in this argument: first,
it is not clear why tobacco-related health services should be financed by a specific
insurance scheme instead of a general one covering all health services. For ex-
ample, Egypt imposes a tax of EGP 0,10 per pack of cigarettes to finance part of a
health insurance programme rather than earmarking a specific insurance scheme
for tobacco-attributable diseases. Second, health spending under this scheme
would have to be narrowly defined; it would exclude, for example, spending
on smoking prevention.
Consequently, earmarking or dedicating revenues from tobacco taxes for the
health system could make more sense. Revenues from tobacco taxes can be sub-
stantial in a number of countries and can provide important resources for health,
particularly in low income countries where resources are scarce. WHO estimates
show that current revenues (2008 data) from excise taxes can represent more than
50% of government health expenditures in countries like Democratic Republic
of Congo, Pakistan or Viet Nam. Even dedicating the resulting revenues of tax
increases for health programmes is an efficient way of raising resources inter-
nally, addressing at the same time any political opposition to such tax increases.
A 50% excise tax increase would increase the excise tax revenues of 22 low-income
countries (for which data was available) by 33%. The extra revenue alone would
be equivalent to 29% of these countrys public health expenditures. Revenues
from tobacco excise taxes where consumption is very high are sometimes almost
equivalent to what is spent on health by the government. In 2008, cigarette excise
tax revenues generated by a 50% excise tax increase were equivalent to 31% and
26% of government health expenditures in Pakistan and Viet Nam respectively
(WHO, 2010).
Tobacco taxes are earmarked by a number of governments. For instance, several
US states (notably California, Massachusetts, Arizona, and Oregon) and several
countries (e.g. Ecuador, Egypt, Estonia, Finland, Iceland, India, Korea, Nepal, and
Thailand) earmark part or all their tobacco tax revenues for different purposes.
In the case of health programmes, these include mainly tobacco control and/or
health promotion. Earmarking tobacco taxes for health purposes is practiced by
 THE POLITICAL ECONOMY OF TOBACCO TAXATION 101

more than 20 countries around the world (WHO, 2009c). In California, 57% of
the excise tax funds the Children and Families First Trust Fund, 29% is spent on
health education, hospital services, physician services and research, and another
2% of the excise funds the Breast Cancer Fund. In the light of the success of an
earmarked tobacco tax in California, similar earmarking of part of the state excise
on cigarettes also takes place in Kentucky (mainly on cancer research), Louisiana
(primarily for tobacco prevention), Massachusetts (mainly on health insurance)
and Oregon (mainly for the health fund). Studies from California found, for
example, that cigarette consumption has been reduced as a result of increases in
both taxes and tobacco-control activities funded by the tax increase (Flewelling
et al., 1992; Keeler et al., 1996).
Nepal imposes a 2 paisa health tax per manufactured cigarette (domesti-
cally produced or imported). The revenue generated by this tax is earmarked
for cancer control. Other types of funded activities include social and health
programmes (Argentina, Costa Rica, Jamaica, Panama, Mongolia, Philippines),
programmes for the protection of children, the elderly and disabled popu-
lations (Costa Rica), education (Costa Rica, Iceland, Korea), emergency care
(El Salvador, Paraguay), and sports activities (Colombia, Estonia and to some
extent Switzerland). Several Australian states and New Zealand use tobacco
tax revenues to fund sporting and artistic events that were previously funded
by the tobacco industry.
Thailand may be the best success story to be noted in the case of tobacco
(and alcohol) tax earmarking. In 2001, the Government of Thailand passed the
Health Promotion Foundation Act, which led to the setting-up of the ThaiHealth
Promotion Foundation. ThaiHealth receives 2% of the total national tax revenue
on alcohol and tobacco products equivalent to about US$35 million per year.
ThaiHealth acts as a catalyst and supports groups and organizations that are
already working on public health issues. It reports directly to the cabinet and
parliament each year. The success of ThaiHealth has inspired other countries to
adopt or contemplate setting up the same policy. For example, Mongolia and Togo
have adopted the same structure as Thai Health and received technical assistance
by ThaiHealth in the process of setting up the policy.
Annex Table 5 summarizes tobacco tax revenue earmarking in various countries
at the central and sub-central levels of government. As one would expect, the link
between revenue and spending is weak, with only a portion of tobacco revenue
earmarked to spending programmes in the majority of countries. For example,
of the 53 countries currently in the WHOs European region, 9 of them earmark
taxes for tobacco control and other public health measures; the average level of
102 Chapter IV

allocation is less than 5 percent of total tax revenue (WHO, 2009c). Moreover,
these programmes tend to be broadly defined, for example,health, education,
social security. Earmarked funds that support broad health and social servic-
es (such as other disease programmes) broaden the political and civil society
support base for tobacco control. For example, in Australia, historically, broad
political support from the Ministries of Sports and Education helped convince
the Ministry of Finance that raising tobacco taxes was possible. Indeed, after
earmarked taxes passed, the Ministry of Finance went on to raise tobacco taxes
further without earmarking (Galbally, 1997). Only a small number of countries
earmark revenues to tobacco control activities and cancer treatment, which could
be considered as narrowly defined spending programmes.
Additionally, targeting revenue from tobacco taxes to other health programmes
for the poorest socioeconomic groups could produce double health gainsreduced
tobacco consumption combined with increased access to and use of health servic-
es. In China, a 10 percent increase in cigarette taxes would decrease consumption
by 5 percent and would increase government revenue by 5 percent. The increased
earnings could finance a package of essential health services for one-third of
Chinas poorest 100 million citizens in 1990 (Saxenian and McGreevey, 1996).
For countries, particularly low and middle income countries where health
coverage is low, tobacco excise tax revenues earmarked or dedicated, depend-
ing on political support can provide an important source for much needed
expenditure on health.