Fat tax lax facts

By Bill Kaye-Blake 02/08/2013 5


Today’s fat tax news:

A fat tax is being suggested as a possible way to improve the health of New Zealanders by encouraging people to replace some saturated fats with polyunsaturated fats.

Some butter could be replaced with grapeseed or safflower oil, meat with omega-3 rich fish, and potato chips with nuts or seeds, while the tax could add $1 to a pack of butter.

Rachel Foster and Associate Professor Nick Wilson from Otago University in Wellington looked at five meta-analyses to estimate how the risk of cardiovascular disease could be reduced by eating less saturated fats.

They also used data from the New Zealand Adult Nutrition Survey 2008/09 to determine whether a change to the amount of fat eaten would be warranted and feasible.

Their conclusion was that replacing 5 per cent of daily energy consumed as saturated fat with polyunsaturated fats would be expected to reduce cardiovascular events by about 10 per cent.

This news report, and others on the newly published article from the Australian and New Zealand Journal of Public Health (paywall), makes two statements:

  • the tax adds $1 to a pack of butter
  • changing the fats would reduce cardiovascular (CV) events by 10%.

Now, if you aren’t careful, you might think these two statements are related. They aren’t. The article, by the way, is the same. It says that a Danish-style tax would add $1 to a 500-g pack of butter. Then, later on, it focuses on switching out the type of fat for 5% of energy consumed.

Let’s connect those two statements.

One way is to go from prices to impacts. Assuming everything works mechanistically, what is the CV events effect of a Danish-style tax?

  • Price of butter: $5 (range on-line: $4-$6)
  • Tax: $1
  • Proportional change in price: 20%
  • Elasticity: assume -0.5 (see here, Table 1)
  • Proportional change in butter consumed: -10%
  • Energy as saturated fats: 13% (from Foster and Wilson)
  • Reduction in energy from saturated fats: 1.3%
  • CV events effect: 2.6%.

Yes, this assumes that the decreased butter is replaced by the required poly-unsaturated fats, but I’m assuming that for the sake of argument. My point is that the $1 tax wouldn’t lead to a 10% drop in CV events, but something closer to 2.6%, even if it worked as advertised.

We can also do the exercise in the other direction:

  • CV events effect: 10%
  • Change in butter consumed: -38% (that is, 5%/13%)
  • Tax: 77%
  • Price of butter: $8.85.

There’s a headline for you: Public Health Academics Propose $9 Butter.

UPDATE: The Manawatu Standard talked to me about the Foster and Wilson article, and shared my estimate of the price impact.


5 Responses to “Fat tax lax facts”

  • Hi Bill – good to see an economist engaging with the public discourse on the food taxes topic. But perhaps you didn’t actually read our article – as we didn’t link the price of butter (post-tax) with any estimate of its reduced consumption. This impression might have been given by poorly written media reports – but not by us in our article. What we actually did was to examine the results of the meta-analyses and conclude that a 5% shift in dietary energy from saturated fats to polyunsaturated fats would reduce the risk of cardiovascular events by around 10% (http://onlinelibrary.wiley.com/doi/10.1111/1753-6405.12080/abstract).

    We did as a matter of context mention that a Danish style saturated fat tax would result in a $1 increase in the price of 500g of butter in NZ – but we didn’t calculate what impact this would have. If there was indeed a $1 increase in the price of butter – your maths are perfectly correct (for the price elasticity you assume). But this is not what we were trying to estimate.

    Nevertheless, a recent article reported that the Danish saturated tax did reduce the consumption of oils and fats by 10-15% (by Danish economists Jensen and Smed in the peer-reviewed journal “Food Policy” – I can supply a copy if you email me: nick.wilson@otago.ac.nz). This was after a 22% price increase for butter and 8% increase for vegetable oils (so Bill you guessed well for your price elasticity of 0.5 for butter). There are complexities with this though, eg, the stockpiling of foods like butter prior to the tax coming into force and only.

    Discussion on these issues is good – but it does help that it is informed by actually reading what people write before starting to blog.

    Regards Nick Wilson (co-author of the article being discussed)

  • Hi Nick – Thanks for commenting. As it happens, I did read the journal article before blogging. I do realise the that Danish tax comment is in the introduction, while the 5%/10% calculation is the meat of the article. That is, the two are technically unrelated.

    The problem is partly on the reporting side of things. For that reason, this blog post actually focuses on the news reporting. The articles on the paper tended to follow a similar pattern: mention $1 a pack for butter then discuss your findings. So, we got juxtapositions like the Stuff article (#1 on Google for ‘fat tax nz’). The headline says ‘Fat tax would add $1 to butter’ and the article starts ‘A fat tax ….’ But which fat tax? The $1 tax? No, that is not the tax that will drop cardiac event by 10%, but nowhere is that made clear.

    However, this is also a problem of science communication (so it is good that we are having this discussion here). I think the Royal Society would agree that we need to improve the quality of science communication in NZ, and that includes ensuring that people don’t think a 20% tax on butter would drop heart attacks by 10%. In the journal article, you and your co-author discuss possible interventions that could help achieve the 5% change, but do not explicitly state that the Danish tax would not be sufficient. In fact, you raise the ‘Danish-style tax’ in the discussion section (p. 335a). I would suggest, given the way the story has been reported, that a statement of its lack of sufficiency could have been helpful.

    Finally, you may be aware that I have worked with Tony and Cliona (we have a paper under review). I or one of my many food economics colleagues would have been happy to help with exactly this sort of back-of-the-envelope calculation. Thus, rather than the last section of your article noting that the impact depends on tax rates, pass-throughs, etc., we could have helped add some indication of the size of the intervention required.

    It is exactly this kind of conversation across economics and public health that can lead to better policy, so thanks for contributing.

    Regards, Bill

  • Hi Bill – thanks for your comment, and very thoughtful too. You are right that the media play fast and loose and such things and science communication by the media needs to improve.

    We did not think it was warranted to estimate what a Danish style tax would do in NZ – as it is just part of the mix since the general trend is for butter to become more expensive than other fats and oils anyway (part of the global market NZ is in). Also New Zealanders are generally switching to greater consumption of vegetable oils for various health and culinary reasons.

    It is also certainly very good for public health people to team up with economists so that there is well considered information for public discussions and policy-maker considerations. So we might well work together next time.

    Best wishes, Nick

  • Hi Bill and Nick,

    I am no economics expert so my contribution does not follow exactly with your comments above. That said I have some points to make, particularly with regard to the relationship between science and the media and the social impacts of the tendency to think uncritically about consumer behavior.

    First as scientists (social or otherwise) it is incumbent on us to think seriously about how the media may or may not understand what our research says. Nick – I read your original article as a first point of reference after I noticed the stuff article. I agree entirely with Bill here – the references to the ‘Danish experiment’ may be used contextually, but they are one of the only accessible handles a non-scientist could use to make sense of your findings. The resultant media version is therefore perhaps a little predictable? An alternative would have been to reframe the article along the lines of what would be ‘actually required’ to achieve the aim you set. This way the general public would not be left with the opinion that increasing the price of butter by $1 will lead to such a dramatic reduction in CV events.

    Misunderstanding science is a preoccupation of the general public. Research doesn’t ‘land’ in the social in the way that the scientists and policy makers intended – it is a constant source of frustration and confusion. Good science communicators is obviously one way to address the issue (my wife is one of those!) but to me the answer also extends the other way – that is we need to make a serious attempt to train scientists to understand the social in a critical way. To recognize for instance the impact of the unconscious, and the way the ‘healthy’ discourse and the business of health (weight loss industry for instance) impact on how people make sense of research. I think this would give scientists more skills to understand how throw-away mentions become the take-home messages.

    All that said – this engagement is fantastic, we now need a media representative to also comment here – I’ll work on that!

    Dr Andrew Dickson,
    School of Management, Massey University.