Monday, 6 October 2014

Sugar tax

I was on the CBC's The 180 on Sunday, discussing sugar taxes, fat taxes and paternalism. The audio's at the CBC site.

While lots of people can point to expanding waistlines as something they find undesirable, just not liking how other people look isn't a sound basis for public policy.

To the extent that costs are borne by the public health system, it's not even obvious whether sugar-linked obesity increases or decreases overall health care costs: there's higher cost at every age, but earlier death. Which effect dominates really isn't clear.

Further, much of the effect is what economists call a transfer. Sure, there will be some effect on total consumption when the cost of the medical bill is externalised, but that's different from the costs of being unhealthy. Most of the costs of being unhealthy are borne by the person who's sick. Maybe you'll consume a bit less exercise and a bit more unhealthy stuff, at the margin, when your health insurance bill doesn't vary with your decisions, but you still have to live with the consequences. People would take less care when driving if their insurance premiums weren't affected by the number of speeding tickets they get, but they'd hardly start rolling their cars on purpose.

For the part of the effect that is due to changes in behaviour under a public health system, implementing a full tax-and-subsidy mechanism to make people behave as they would if they were under an insurance system winds up undoing much of the benefit of a public health system: it replicates insurance premiums, but at a higher administrative cost.

Finally, there's just something wrong with telling people not only that they have to be signed up for a compulsory insurance system that they might not want, but also that because they're signed up for it, they're going to be subject to a pile of taxes and subsidies and regulations to make sure that the overall system isn't too expensive.

And even if all of that could be swept aside, there isn't particularly good evidence that soda taxes even work. Instead, there's just substitution over to calories from non-soda drinks. A comprehensive, and very high, sugar tax would change behaviour. Tobacco taxes have reduced smoking considerably. But recall that, at least in New Zealand, you're paying $0.55 in tax per cigarette. If the average price of a pack of 20 cigarettes is about $17.20, then tax is 63% of the cost of a pack. A comprehensive tax on all sugars that resulted in a $0.50 increase in the cost of a can of Coke might also reduce sugar consumption. There's no particularly good reason for doing such a thing, or at least not one motivated in things other than your aesthetic preferences over how other people should behave, but that high a tax could have effects.

Paternalistic regulations on consumption are insulting. They're classist in application. When they're based on "costs through the public health system", they are unlimited in potential range. And when they're packaged up as "let's tax soda and subsidise healthy foods", you empower an army of rent-seekers to argue over the edge cases.


  1. Christian BjørnskovMon Oct 06, 08:40:00 pm GMT+13

    Denmark introduced a fat tax at the beginning of 2012 and abandoned it again a year later. Besides the problems that Eric outline, the fat tax turned out to be a bureaucratic nightmare. Firms had to somehow assessed the fat contents in whatever they produced, which turned out to be both cumbersome and very expensive. A producer of high-quality sausages in Southern Denmark had to reduce their product line because about half of their products were unprofitable with the transaction costs associated with documenting their fat contents.
    The introduction of the fat tax came to be an example of a problem that is almost always ignored by politicians: that any public authority with little accountability to voters is likely to shift as many transaction costs unto the private sector and voters as possible. Furthermore, the media wrote about the decision of how to reach the 'correct' tax rate. It turned out to be calculated as what was needed to yield 4 billion kroner (about .9 billion NZD).

  2. Here's a supplementary point I think you'll agree with: When priorities are set on the basis of a metric such as life-years lost, one will miss an important difference between communicable and non-communicable diseases: When doing the things that typically lead to non-communicable diseases, people often get something in return - e. g., the buzz you get after the third glass of beer. This is not the case with catching the flu or ebola.

  3. That was a terrible interview. "people shouldnt have sex because they may get an std" seriously????? Thats a terrible argument. Foolish.

  4. If the costs of behaviour to the public health system are the reason for banning or taxing an activity, then there really isn't much that wouldn't be subject to tax and regulation, sex included.

    But we don't have to say that costs to the health system justify paternalistic regulation. That's the much simpler, and better, answer.

  5. Exactly. Addressing communicable diseases works to solve a market failure: pandemic protection is a strong public good.