Sunday, January 8, 2017

Public Health Strategies Part 4: Taxes

In an earlier post I mentioned a few different commonly proposed public health strategies, and how most of them fit quite nicely into a political spectrum characterised by an authoritarian-libertarian axis and a left-right axis.   In this post I’ll discuss the strategy of taxation/subsidies, which I thought fits nicely into the authoritarian left quadrant.  This is because this strategy holds the food environment primarily responsible for people not adopting healthy lifestyle behaviours rather than personal responsibility, and then uses government controls to manipulate the free market


There are a few targets people have suggested to tax or subsidise.  There was the failed Danish tax on saturated fat, and sometimes there are suggestions that fruit + vegetables and even gym memberships (because you couldn’t possibly exercise without one) should be subsidised, but this isn’t very common.  At the moment, a tax on sugar sweetened beverages (SSBs) is far more commonly proposed and is actually being implemented in a few areas, so I’ll focus on that

There are two main rationales to support a tax on SSBs.  The first is that a tax on SSBs will increase their price and this would reduce the consumption of them, particularly in people on lower incomes, who consume more SSBs on average, as they are more sensitive to changes in price.  This is textbook supply and demand, as price goes up demand decreases.  In this respect, a tax on unhealthy food or a subsidy on healthy food could be seen as a form of nanny statism.  Generally, implicit in the support for a tax on SSBs is the belief that education programs haven’t or will not work on some people (the laggards).  This usually isn’t mentioned when a tax on SSBs is proposed (probably out of political correctness and not risking alienating others), but it is an important premise to justify taxes/subsidies, because otherwise why not use education instead?  And so a tax on SSBs appears necessary to reduce SSB consumption and therefore improve health, or, to save some people from themselves.  Or perhaps more importantly, to save the overweight/obese children with a mouthful of tooth decay from negligent parents.  But people who are against many forms of nanny statism argue that what right does the government have to control/influence individual choices that don’t adversely affect others?*  In addition, cigarettes are heavily taxed but plenty of people still smoke, and as David Gillespie points out, there is already a tax in Australia on many processed foods including SSBs in the form of the GST [1]

This leads into the second rationale for taxing SSBs which counters the previous argument.  In countries with a socialised healthcare system the consumption of SSBs creates a negative externality.  When people purchase SSBs they are paying the costs for the product but are mostly or totally externalising the increase in healthcare costs from SSB consumption increasing the risk of several chronic diseases.  As a result, it could be said that they are not paying the ‘true costs’ of a SSB.  Let’s say 1 litre of SSBs costs the consumer $1.50 but each litre of SSBs is associated with an increase in healthcare costs at about $0.30.  Therefore, in this scenario a 20% tax on SSBs is necessary to internalise the healthcare costs.  With this rationale of internalising externalities, a tax on SSBs could be considered a successful policy even if no one changes their behaviour as a result of the tax, and also doesn’t make it about nanny statism or a moral judgement (a ‘sin tax’)

Of course, the government (and health associations [2]) also has a motivation to tax things that are politically acceptable as a form of cash grab, particularly if they’re the kind of government that likes to spend money, which seems to be the main motivation for the Danish tax on saturated fat [3].  “A lesson learnt from this chain of events is that if a tax on fat is to survive it needs more than merely to be passed. It probably needs to be politically supported for health rather than fiscal reasons and to be supported or at least accepted by prominent actors in the food arena including researchers.” [3]

* This can also apply to laws against recreational drug use (that doesn’t expose people to second hand smoke for example), ‘unsafe’ playgrounds and bike helmet and seatbelt laws


Danish tax on saturated fat: some people in public health have praised the Danish fat tax because it reduced the intake of saturated fat [3].  However, simply implementing a well-intentioned policy doesn’t mean it’s necessarily going to have good outcomes.  The most important outcome of such policies should be related to health rather than consumption, and in this respect even diet-heart diehards should judge it to be a failure.  Based on ecological data, the Danish tax on saturated fat appears to have reduced SFA, MUFA and PUFA by 0.3%, 0.2% and 0.1% of total calories respectively (this is because foods high in SFA are often high in MUFA and PUFA on a grams per weight basis).  As a result, LDL-C would be expected to decrease by 0.008 mmol/l and HDL-C would be expected to decrease by 0.005 mmol/l, and this would be expected to increase the risk of CHD by 0.2% (-0.3% for LDL-C +0.5% for HDL-C) [4].  While it’s ideal to be able to judge the efficacy of the tax based on actual changes in population health rather than modelling, the estimated effect size here is probably too small to notice and would likely be drowned out by noise.  The Danish fat tax also had the issue where consumers could avoid the tax by purchasing heavily taxed foods in neighbouring countries without such a tax

SSB tax in Australia: a tax on SSBs hasn’t been implemented in Australia but one study in particular has estimated effect of a 20% tax on SSBs in Australia, and found the following [5]:

141g/d to 124g/d
76g/d to 67g/d
Total energy intake
-16 kJ/d
-9 kJ/d
Change in BMI
Weight loss
0.32 kg
0.06 kg
Obesity prevalence
2.7% (0.7 percentage points)
1.2% (0.3 percentage points)
Health-adjusted life years
112,000 (4.54 d/per capita)
56,000 (2.27 d/per capita)

The estimated change in BMI as result of the tax is similar in magnitude to other studies in the US [6].  The tax is expected to cost 27.6 million AUD upfront, but generate 400 million each year and reduce healthcare costs by up to 29 million per year (savings are expected to increase over time and plateau at 29 million) [5].  This is a decrease of just ~0.024% in total healthcare expenditure (29 million/121.4 billion [7]) and is why it’s important to put those kinds of figures in context

Given that the revenue generated equates to about 0.329% of total healthcare expenditure, and that SSB consumption is likely responsible for at least a 1% of total healthcare expenditure (not aware of an estimation for this, but 1% seems like a reasonably conservative estimate), if the rationale for taxing SSBs is to internalise externalities then the tax would need to be substantially higher.  This is also true if the goal is to meaningfully improve population health

One of the problems with this model is that the health outcomes are based on BMI, which in turn are based on the very small estimated reduction in calorie intake, and doesn’t look at any effects of SSB intake independent of calorie intake and BMI.  And of course it looks at adults, whereas if it also included children the estimated effects would be greater because the analysis would include more people and because the expected health benefits are larger for younger people.  That being said, a modest (~20%) tax on SSBs in an affluent country like Australia (where the economic effect can be largely ignored by almost everyone) is certainly not going to be a silver bullet.  The food industry has a point when they say that SSBs only contribute about 3% on average to total energy intake.  ‘Extra’ foods contribute about 36% on average to total energy intake in Australia [8], so there’s a lot wrong with the average Australian diet.  Therefore, one could argue a lot of potential targets to tax, but also that there’s a lot more that needs to be changed by even a 10% reduction (~3 percentage points) in ‘extra’ foods as a result of a broad taxation policy

Other Objections

Some people argue that a tax on SSBs is a tax on the poor, as poorer people on average consume more SSB and are more motivated by changes in price, but there are a few issues with this objection.  (1) They are confusing intent with outcome.  This is a mistake many people make related to other issues where there is a race/sex/etc disparity.  The tax is not discriminatory (except against SSBs) as it’s not intended to disproportionately tax poorer people, but it’s simply that the people who consume more SSBs happen to be poorer.  It’s not like this is a tax on renting or living in apartments, which actually has more of a causal relationship with being poorer (2) People are free to purchase SSB or not.  There’s nothing forcing poorer people to consume more SSBs on average.  If they don’t want to pay the tax they can choose not to purchase them, after all SSBs are very discretionary food items.  (3) While the tax is estimated to cause large relative differences in total household expenditure (consumption multiplied by income differences), the actual increase in expenditure for low income earners is very modest (~0.2% household income, ~1% of food budget) [9] (4) A tax on SSBs is likely to disproportionately benefit the health of poorer people.  People in public health usually try to engage and improve the health of low SES people disproportionately.  After all, no one in public health is really advocating for a subsidy on salmon, avocado, blueberries and quinoa.  (5) Who do you suppose the tax revenue is likely to disproportionately benefit, almost regardless of what it’s used for?

Another objection is that a tax on SSBs will reduce sales and as a consequence some employees will have to be laid off (if the food industry objects to the tax for this reason, then you know they think it will work to some extent).  The goal isn’t (and shouldn’t be) to punish the food industry as they are reacting to consumer demand for the most part (but they should be punished when they distort science, and there are numerous examples of this).  But the food industry and their employees shouldn’t get upset that in countries with socialised healthcare, the government at some point may tax unhealthy food to improve health and/or internalise externalities, after all there is a rationale and incentive for it to do so.  While it’s unfortunate that a very small number of people may lose their jobs, it’s important to recognise that there are other jobs out there, and that changes in the world will always create winners and losers (the internet is a great example), and part of life is setting yourself up to manage such likely changes well

While I thought taxes/subsidies and bans fit nicely into the authoritarian left quadrant, a modest (~10-20%) tax on SSBs isn’t really that authoritarian.  However, a major concern, particularly from libertarians, is that a tax on SSBs sets up a precedent for governments to tax other things, where these other things may be quite inappropriate targets for improving population health, while also leading to a progressive loss in individual freedom.  This should also be of concern to people in ancestral health, low carbers and other, regardless of political opinion, as by adopting such dietary practices we acknowledge that the government and mainstream isn’t always right.  A good example of an inappropriate target is the Danish tax on saturated fat, as saturated fat is not associated with coronary heart disease in meta-analyses of observational studies [10], replacing saturated fat with polyunsaturated fat does not reduce coronary heart disease in adequately controlled randomised controlled trials [11], and reducing fat intake is hardly the best strategy for weight loss [12].  Similarly, total fat is a potential target, but increasingly less likely.  (Red) meat is another potential target for taxation for health or environmental reasons, despite evidence to the contrary [13] [14] [15].  Salt is another potential target because people in public health seem to have an almost pathological hatred of salt, despite the relationship between salt intake and mortality being on a U-shaped curve [16] and that reducing salt intake doesn’t affect blood pressure much but does have some undesirable side-effects [17].  Being in academia and around nutrition students hasn’t alleviated these concerns, it has strengthened them

This post is already quite long so I’ll simply list a few more papers if you’re interested:

  • Beverage purchases from stores in Mexico under the excise tax on sugar sweetened beverages: observational study [18]
  • Impact of the Berkeley Excise Tax on Sugar-Sweetened Beverage Consumption [19]
  • Using price policies to promote healthier diets [20]
  • Modelling the potential impact of a sugar-sweetened beverage tax on stroke mortality, costs and health-adjusted life years in South Africa [21]

And finally, if you want to follow me on Facebook you now can.  Just follow this link​ and like the page