Appendix A Sugar taxes around the world
Table 9 Sugar taxes around the world
Jurisdiction
Year
Taxed products
Tax design
Source
implemented
Albania
2019
Beverages with
(10 leh per litre – around
Obesity
more than 5g of
8%)
Evidence Hub
sugar per 100ml
n.d.
Albany,
2017
Volumetric per ounce of
Obesity
California,
beverage
Evidence Hub
U.S.
n.d.
Bahrain
2017
Aerated
Excise tax – 100% tax on
World Action on
flavoured soft
energy drinks, 50% tax on
Salt, Sugar and
drinks, and
soft drinks
Health n.d.
concentrates,
powders, gels or
extracts, which
will be made
into an aerated
drink
Barbados
2015
All sugar-
10% excise tax
Obesity
sweetened
Evidence Hub
beverages
n.d.
Belgium
2016
Beverages with
Specific
World Health
added sugar,
(volumetric), flat
Organization.
sweeteners or
Regional Office
rate (product type)
flavours
for Europe 2022
6.8133€/hl
Bermuda
2019
Sugar-
75% on the customs value
Obesity
sweetened
Evidence Hub
beverages
n.d.
Berkeley,
2015
Sugar-
Volumetric excise tax (1 cent
Obesity
California,
sweetened
per ounce of beverage)
Evidence Hub
under the Official Information Act 1982
U.S.6
beverages
n.d.
excluding milk-
based
Boulder,
2017
Sugar-
Volumetric excise tax (2
Cawley et al.
Colorado,
sweetened
cent/ounce on drinks with >
2021
U.S.
beverages with
5g caloric
at least 5 grams
sweetener/12ounces)
Released
of caloric
sweetener per
12 fluid ounces.
British
2021
Sugar-
Removal of exemption from
Government of
Columbia,
sweetened and
provincial sales tax – 7%
British
Canada
artificially-
sales tax applied.
Columbia,
sweetened
Ministry of
beverages
Finance (2021)
6
The state of California decided in 2018 to ban cities from introducing any further taxes on SSBs until 2030 (Schmacker and Smed,
2020).
89
Jurisdiction
Year
Taxed products
Tax design
Source
implemented
Brunei
2017
Beverages with
$0.02 per litre (about 2%)
Obesity
more than 6
Evidence Hub
grams of sugar
n.d.
per 100 ml
Catalonia,
2016
Sugar-
8 cent/litre on drinks with 5–
Griffiths et al.
Spain
sweetened
8g sugar/100ml,
2019
beverages
12 cent/litre on drinks with
>8g sugar/100ml
Chile
2014
Sugar-
Ad valorem 10% on drinks <
Griffiths et al.
sweetened
6.25g
2019
beverages
sugar/100ml,
18% on
1982
drinks > 6.25g of
sugar/100ml
Act
Cook
2017,
Sugar-
The county passed a
(Urban Institute
County,
repealed
sweetened and
volumetric excise tax (1 cent
n.d.)
Illinois, U.S.
same year
artificially-
per ounce) tax in November
(includes
sweetened
2016. The county repealed it
Chicago)
beverages, but
in October 2017.
excluding those
purchased with
food stamps and
fruit drinks
Information
Cook
2016
Sugar-
$0.30 per litre (about 30%)
Obesity
Islands
sweetened
Evidence Hub
beverages
n.d.
Denmark
1930s –
soft drinks (at
Per litre of beverage
Obesity
increased in
various times,
Evidence Hub
Official
January 2012,
other sugary
n.d.
decreased in
products and
Schmacker and
the
July 2013,
also products
Smed, 2020
completely
high in fat have
repealed in
also been
2014.
subject to taxes)
Domenica
2019
Sugar-
$0.01 per gram of sugar
Obesity
under sweetened
content above 5 grams per
Evidence Hub
beverages
serving
n.d.
Estonia
2018
Sugar-
€0.10 per litre (about 10%)
Obesity
sweetened
on beverages with more
Evidence Hub
beverages
than 5 grams of sugar per
n.d.
100 ml
Released
Fiji
2018
Sugar-
$0.30 per litre (about 30%)
Obesity
sweetened
on drinks with more than 6
Evidence Hub
beverages
grams of sugar per 100 ml
n.d.
Finland
2011/2014
Sugar-
Volumetric specific excise
World Health
sweetened
tax 75 cents/kg or 7.5 cents/l
Organization.
beverages, fruit
for liquids, 75 cents/kg for
Regional Office
and vegetable
solid ingredients of soft
for Europe 2022
juices
drinks
France
2012
Beverages with
Volumetric specific excise
World Health
added sugar or
tax 7.16€/hl (around €0.075
Organization.
sweeteners
per litre, or 7.5%) and an
Regional Office
additional tax of €0.02 per
for Europe 2022
90
Jurisdiction
Year
Taxed products
Tax design
Source
implemented
gram of sugar content above
Obesity
5 grams per serving since
Evidence Hub
2018
n.d.
Hungary
2011
All foods and
Volumetric specific excise
World Health
drinks with high
tax 5 forint/l for >8 g
Organization.
sugar and/or
sugar/100 ml
Regional Office
salt content or
250 forint/l for energy drinks
for Europe 2022
artificially
Obesity
sweetened
Evidence Hub
n.d.
India
2017
Sugar-
Higher GST (18%) on all SSBs
Obesity
sweetened
Evidence Hub
1982
beverages
n.d.
Indonesia
2020
Sugar-
IDR200 per gram of sugar
Obesity
Act
sweetened
content above the threshold
Evidence Hub
beverages
level for different categories
n.d.
of SSBs
Ireland
2018
Sweetened
Volumetric-specific excise
Crosbie et al.
drinks
tax
2022
containing more
€0.20 per litre (about 6%)
than 5g of sugar
per 100ml
Information
Israel
2022
Sweetened
NIS 1,00 per litre for
Israel Tax
(repealed in
beverages,
beverages containing 5g or
Authority 2021
2023)
concentrates
more of sugar per 100ml
and powders
NIS 6,00/litre for
concentrates and NIS
Official 6,00/kilogram for powders,
designed to produce a
beverage containing 5g or
the
more of sugar per 100ml
Italy
2023
Sugar-
€10 per hectolitre for the
Dosen 2022
sweetened and
finished products, and 0.25 €
artificially
per kilogram, for products
under sweetened soft designed to be used after
drinks
dilution
Latvia
2018
Sugar-
€0.075 per litre (about 7.5%)
Obesity
sweetened
on drinks with more than 5
Evidence Hub
beverages
grams of sugar per 100 ml
n.d.
Lithuania
2017
Sugar-
€0.055 per litre (about 5.5%)
Obesity
Released
sweetened
on drinks with more than 5
Evidence Hub
beverages
grams of sugar per 100 ml
n.d.
Maine
2003
Sugar-
5.5% sales tax
Madsen 2020
sweetened and
artificially-
sweetened
beverages and
snack foods
Malaysia
2019
Sugar-
RM0.40/litre (about 10%) on
Pakiam 2019
sweetened and
SSBs with more than 5g of
artificially-
sugar or sweeteners per 100
ml, or 7g per 100ml if milk-
91
Jurisdiction
Year
Taxed products
Tax design
Source
implemented
sweetened
based, 12g per 100ml if
beverages
fruit/vegetable juice based
Mauritius
2016
Sugar-
MUR3 per gram of sugar
Obesity
sweetened
content above the threshold
Evidence Hub
beverages
level for different categories
n.d.
of SSBs
Mexico
2014
Sugar-
Volumetric per litre of
Griffiths et al.
sweetened
beverage
2019
beverages
MXN1 per litre (about 10%)
Morocco
2013
Sugar-
MAD0.70 per litre on soft
Global Food
sweetened and
and non-carbonated drinks
Research
1982
artificially-
with ≥5 g sugar per 100 mL,
Program 2020
sweetened
0.6 MAD/L on energy drinks;
beverages
MAD 0.15/L ($0.02) on
Act
nectars
Nauru
2007
Sugar-
30% import duty
Global Food
sweetened
on all products with added
Research
beverages
sugars
Program 2020
Newfound-
2021
Sugar-
CDN$0.20 per litre,
Obesity
land and
sweetened
volumetric (per litre of
Evidence Hub
Labrador,
beverages
beverage) levied on
n.d.
Canada
wholesalers and/or retailers7
Information
Norway
2017,
Sugary drinks
Volumetric, Per litre of
Øvrebø et al.
increased in
Sugar-
beverage
2020
2018,
sweetened and
NOK3.34 per litre (about
Global Database
repealed in
artificially-
34%) since 1922
on the
Official
2020
sweetened
additional tax of NOK4.75
Implementation
beverages
per kilogram of sugar
of Nutrition
content above the threshold
Action (GINA)
the
level for different categories
n.d.
of SSBs since 2018, but
repealed in 2020
Oakland,
2017
Sugar-
Volumetric excise tax per
Obesity
under
California,
sweetened
ounce of beverage with ≥25
Evidence Hub
U.S.
beverages sold
kilocalories/12 ounces.
n.d.
in fast food
restaurants
Ohio, U.S.
1991
Sugar-
5% increase in sales tax
Obesity
sweetened and
Evidence Hub
artificially-
n.d.
Released
sweetened
beverages
Peru
2018
Sugar-
17% on drinks with more
Obesity
sweetened
than 6 grams of sugar per
Evidence Hub
beverages
100 ml, and a tax of 25% on
n.d.
drinks with more than 8
grams of sugar per 100 ml
7
The provincial Finance Department, has stated that the provincial government collects the tax at the wholesale level, but most
retailers must "levy and collect the sugar-sweetened beverage tax from consumers and pay over the tax amount collected to the
registered wholesaler, as per the wholesaler's invoice."
92
Jurisdiction
Year
Taxed products
Tax design
Source
implemented
Philadelphi
2017
Sugar-
Volumetric excise tax on
Jones et al. 2019
a, U.S.
sweetened and
distributors (1.5 cents per
artificially-
ounce of beverage)
sweetened
beverages
Philippines
2018
Sugar-
Volumetric PHP6 per litre
Saxena et al.
sweetened
(about 12%) on drinks with
2019
beverages
caloric sweeteners, and a tax
of PHP12 per litre (about
24%) on drinks with high-
fructose corn syrup
Portugal
2017
Sugar-
Volumetric specific excise
Gonçalves and
1982
sweetened and
tax 8 cent/litre on drinks
Merenda 2022
artificially-
with < 8g sugar/100ml, 16
Griffiths et al.
sweetened
cent/litre on drinks with > 8g
2019
Act
beverages
sugar/100ml
Qatar
2019
Sugar-
Ad valorem 100% on energy
Koe 2019
sweetened
drinks and 50% on other
beverages
SSBs
Samoa
2017
Sugar-
Volumetric tax on importers
Teng et al. 2021
sweetened
52.5 sene/L (around 21%)
beverages
Information
San
Sugar-
Volumetric excise tax (per
Treasurer & Tax
Francisco,
sweetened
ounce of beverage) of one
Collector 2021
U.S.
beverages,
cent per fluid ounce excise
syrups, and
tax on the initial distribution
powders
within San Francisco
Official
Saudi
2017
Sugar-
Ad valorem 50% on all SSBs
Jalloun and
Arabia
sweetened
in addition to the 2018 VAT
Qurban 2022
beverages
on all consumption goods
the
Seattle,
2018
Sugar-
Volumetric US 1.75 cents per
City of Seattle
Washingto
sweetened soda
ounce of beverage
201)
n, U.S.
and some fruit
drinks
under
Seychelles
2010
Sugar-
Volumetric SCR4 per litre
Global Food
sweetened
(about 40% or USD 0.22) per
Research
beverages
litre import tariff
Program 2020
containing >5 g
sugar/100 mL
South
2018
Sugar-
Specific excise tax per gram
Stacey et al.
Released
Africa
sweetened
of sugar over 4g per 100ml
2019
beverages
ZAR0.021 (approx. 0.15 US
Global Database
cents) per gram of sugar
on the
content above 4 grams per
Implementation
100 ml (about 11%)
of Nutrition
Revenue “soft-earmarked”
Action (GINA)
for health promotion
2018
activities
Sri Lanka
2018
Sugar-
LKR 12/litre or 40
Institute of
sweetened
cents/gram of sugar content
Policy Studies of
beverages with
above the threshold level for
Sri Lanka n.d.
different categories of SSBs
93
Jurisdiction
Year
Taxed products
Tax design
Source
implemented
more than 4 g
sugar per 100 ml
Thailand
2017
Sugar-
Tiered tax based on sugar
Obesity
sweetened
content and product
Evidence Hub
beverages
category, ranging from
n.d.
THB0.13 to THB1 per litre
(about 1-10%)
Tonga
2013
Sugar-
T$0.50/L
World Bank
sweetened
beverages
United
2017
Sweetened
50% sales tax on carbonated
Obesity
Arab
carbonated
beverages, 100% sales tax on
Evidence Hub
1982
Emirates
beverages and
energy drinks
n.d.
energy drinks
Act
United
2018
Sugar-
Volumetric excise tax tiered
HM Revenue
Kingdom
sweetened
by sugar density (18p/litre
and Customs
beverages
on drinks with 5–8g sugar/
n.d.
excluding milk-
100ml, 24p/litre on drinks
based and
with >8g sugar/100ml)
beverages made
by small
producers
Information
Washingto
2010,
Sugar-
Volumetric excise tax (1/6
Griffiths et al.
n State,
repealed
sweetened and
cent per ounce of beverage)
2019
U.S.
same year
artificially-
sweetened
carbonated
beverages
Official
Vanuatu
2016
Sugar-
VUV30 per litre (about 30%)
Obesity
sweetened
Evidence Hub
the
beverages
n.d.
Source: NZIER
under
Released
94
Appendix B Empirical methods for analysing real-world sugar tax
impacts
Table 10 Empirical methods for analysing real-world sugar tax impacts
Method
Contexts
Strengths and limitations
Uncontrolled
The least statistically sophisticated method in
The simplest way to evaluate changes that
pre-post
included studies, and sometimes called before-
occur after the implementation of a tax is to
studies
and-after studies, these studies involve
compare prices and purchases before the tax is
observing a single group of individuals before
implemented with those after the tax is
and after an intervention or policy change, and
implemented. This assumes that in the
measuring the outcome variable at both time
absence of the tax, prices and purchases
1982
points. The analysis typically involves comparing
would have been identical to their levels prior
the means or proportions of the outcome
to the introduction of the tax. This is unlikely
Act
variable before and after the intervention or
to be true for many reasons; for example,
policy change.
seasonal variation in purchases, trends in how
The pre-post approach is typically most
much people like soft drinks, or concern about
convincing when it is applied to a narrow time
the consequences of soft drinks for health are
window around the implementation of the tax;
all likely to lead to changes in purchases, even
in this case, it is more plausible that changes in
if the tax was not introduced.
price and purchases are primarily driven by the
In general, an uncontrolled pre-post study can
tax and less appropriate for ascertaining the long offer robust insights into changes that occur
term impacts of a tax.
after a tax is implemented, but cannot causally
Information
attribute those changes to the tax.
Difference-
Difference-in-differences (DiD) approaches are
DiD assumes that the treatment and control
in-
applied when one or several groups are exposed
groups would have followed the same trend
differences
to an intervention or policy change and others
without the intervention or policy change,
are not. The DiD approach is based on
which may not always be realistic.
Official
demonstrating a stable and predictable
In a DiD study, the treatment and control
difference between the groups in the pre-
groups may differ in important ways, which
intervention period and a change or disruption
the
can introduce selection bias. In particular,
in the difference in the post-intervention period.
when different countries are used for DiD
In the case of sugar taxes, the control group can
groups, they are less likely to be culturally and
be a country, a city, or a product that was not
politically similar than two nearby cities might
subject to the tax.
be, making this approach most relevant in US
under
In the US, where taxes have been introduced
studies of sugar taxes. However, one concern
city- or state-wide as opposed to nationwide,
with using nearby geographical areas as a
the set of nearby areas without a tax provides a
control group is that they may experience
plausible set of control groups.
spillover effects. If residents of the taxed area
decide to shop more in neighbouring untaxed
areas, then this would lead to an increase in
purchases in that area and therefore lead to
the conclusion that the tax led purchases to
Released
fall by more than they actually did. One
approach to deal with this challenge is to
exclude areas immediately adjacent to the
taxed area from the control group.
Controlled
Similar to DiD approaches, studies that use a
CITS can often detect the immediate effects of
interrupted
controlled interrupted time series (CITS)
an intervention or policy change by comparing
time series
approach study the changes in a population or
the outcome variable before and after the
variable of interest over time, including before
intervention. In contrast, DiD may require a lag
and after the intervention or policy change
period before the treatment effect can be
occurs. The outcome variable is measured
observed.
repeatedly before and after the intervention,
Unlike before-an-after studies, CITS studies
and the interrupted time series analysis
can control for pre-existing trends and
95
Method
Contexts
Strengths and limitations
compares the trend of the outcome variable
therefore allow for more robust analysis over a
before and after the intervention to the control
longer period of time.
series which is usually a predicted
A key advantage of this technique over DiD
counterfactual constructed from pre-
techniques is that the requirement to identify
intervention trends.
a parallel trend does not apply here.
CITS is different from DiD due to the focus being
A key disadvantage of this approach is the
on the trend in the outcome variable compared
requirement for sufficient pre-intervention
with the control variable rather than on the
data to construct a predicted counterfactual.
difference between the two.
The quality of the trend analysis and the
predicted counterfactual are reliant on good
quality, complete data.
Other
These methods represent various econometric
These alternative regression-based approaches
regression
modelling approaches where regression
also offer the opportunity for well-controlled
1982
models,
techniques are used to disaggregate changes in
analysis and causal attribution. They are
including
the outcome variable to the various potential
appropriate in any of the same situations
mixed
contributors or causal variables. Commonly,
where controlled-interrupted time series or
Act
effects, time
predicted values for outcomes are generated
DiD approaches may be used, but also where
series
based on the best fit model developed using
data on a control is not available while rich
prediction
pre-tax data. Actual data post-tax is then
data on households, individuals,
models and
compared with predicted values to evaluate the
neighbourhoods and stores, and economic
discontinuity
predictive power of the model and provide
drivers may be available to a range of potential
methods
certainty regarding its causal attribution.
causal factors to be included in the model.
Source: NZIER
Information
Official
the
under
Released
96
