this blog is girtby.net

Junk Food Excise

One of the persistent memes in the Australian media is the notion of childhood obesity caused by poor diet and no exercise. It’s one of those self-evident truths that nevertheless causes much handwringing and below-the-neck stock footage on the evening broadcast.

Recently there was a report released by Access Economics which estimated the cost of obesity (both adult and childhood) at $3.7b. Oops, make that $21b. (From reading the report, it seems the lower figure is from type-2 diabetes alone, and the higher figure from all diseases combined). It’s a lot, anyway.

So what to do? Most discussion has revolved around fairly weak-sounding measures such as removing — or even just toning down — junk food advertising targeted at kiddies.

But I can’t help wondering why an excise wouldn’t help here. We tax cigarettes and alcohol in order to (partly) offset the increased health costs that result from their use. Why should junk food be any different, assuming a suitable definition can be found? Better still, what if the excise were to subsidise not the health costs associated with obesity, but instead subsidise the availability, and price, of less harmful food?

Lynne Pezzullo, who authored the Access Economics report, is sceptical:

For example, she said, there was no point putting a “fat tax” on sugary drinks or fatty foods, because such taxes did not work.

But why not? Tacked onto the end of the Access Economics report is a section (7.4.1) which goes into more detail. It raises many points against the fat tax:

• A fat tax would be an ill-targeted, blunt instrument for improving health outcomes, since it is not targeted at obese people, but rather at food products consumed in some degree by all people. Implicit in this prescription is the assumption that the type of food consumed is a significant contributor to the obesity problem, rather than the amount.

Well, isn’t this a valid assumption? Or are cabbages just as bad for me as Big Macs, in the same amount? This argument is nonsense; of course it is the type of food that is the significant contributor. And if the fat tax can be structured such that the tax payable is directly proportional to the health-related factors in the food, then it will be a specific well-targeted instrument.

This argument seems to be aimed at the straw man that there would be a flat rate per-person License to Eat Junk.

• A fat tax implicitly assumes that higher taxation of such products will substantially shift the pattern of food consumption away from them and towards other, more healthy, foods. However, if the demand for these goods is relatively inelastic (ie, quantities consumed change only in minor ways for any unit price change), then this will not be the case, and the only or main impact will be to increase the price of the product.

That’s funny, I thought it is exactly because of the inelastic demand that we have an excise on alcohol and cigarettes…

Look, I don’t think it’s as simple as saying that people will want junk food no matter what the cost. In my experience, generally junk food is both more available and cheaper than non-junk alternatives. Why is bottled water often more expensive than the Coca-Cola that comes from the same factory? Why is a Big Mac often cheaper than a sandwich? Why would you buy a dodgy-looking apple when there’s a cheaper Mars bar sitting right next to it?

Given that the demand for food is inelastic — a fact conceded in the report with actual references to back it up even — how much of this can be attributed to junk food? Other factors besides cost affect our decision to eat junk. If the relative costs of junk and non-junk can be inverted, surely the demand for the latter will be increased significantly (all else being equal)?

• An example of this principle is that tap water is an essentially free and healthy beverage in Australia. However, its close-to-zero price does not necessarily induce people to drink water rather than (more expensive) carbonated soft drinks. People’s preferences may not be greatly affected by taxation or other interventions to change prices.

This is a very misleading example. Often people drink soft drinks because they are cold. Tap water often isn’t. Anyone who runs a night-club and deliberately disables the cold water taps in the bathrooms knows this.

• Moreover, concessional GST-free status is already extended to certain fresh food while more processed food including most snack food, takeaway and restaurant food is taxed at the normal rate of 10%. There is no evidence that this concessional treatment is effective in switching demand towards GST-free items.

That’s because there is no evidence that the GST concession has resulted in cheaper prices for healthier food, see above.

The next point is basically a repeat of the “inelastic demand” argument:

• Given that demand for food generally is relatively insensitive to price movements, as shown in various econometric analyses (eg, Huang and Lin, 2000; You et al, 1998; Department for Environment, Food and Rural Affairs, 2000); the fat tax is more likely to raise revenue than change consumption patterns.

So you’re saying that it won’t convince people to skip meals or otherwise change their overall eating patterns? Well, duh. “Hmm, I can’t afford that $20 Big Mac, but I’d rather go hungry than eat this $5 tuna sandwich instead!”

• Moreover, a fat tax would be regressive, hurting the poor proportionately more than the rich. Poor people who are not obese would be disadvantaged by the tax, even though their moderate consumption of ‘bad’ foods is not causing them to be obese, particularly if they exercise responsibly also.

But this assumes that the overall food expenditure would increase. In my world, the junk food excise would be used to reduce the price of non-junk food by a corresponding amount. Now of course this might not be realistic for all sorts of reasons. But I haven’t seen them.

• Such a tax would be extremely complex to design and administer, imposing significant compliance costs on the community and administrative costs on the Australian Taxation Office. This was also experienced with the GST, creating numerous anomalies.

Well if you wanted to scare people off a tax on the basis of complexity, citing the GST is certainly a good way to do it. Unfortunately I don’t see any reason why excise cannot be calculated based on existing food categorisation and labelling schemes, such as those cited elsewhere in the report.

Of course it’s not easy, but don’t forget, we have a potential war chest of $21b worth of increased complexity and bureaucracy to draw from!

• Any new public health campaigns would be best funded from the budget (and out of general revenue) rather than through the imposition of a fat tax.

This is just begging the question.

In short, none of these reasons really explain why an excise on junk food wouldn’t work. I say let’s do it. But then again, maybe I’m just feeling superior because I played sport at lunchtime and had a fruit salad after.

Alastair,

Prompted by you, I have read selected parts of the report. I share your concerns that the arguments against “fat tax” are poorly constructed. I am not advocating such a tax, but I don’t think it should be rejected with flimsy arguments like the ones outlined in this report.

First, a clarification regarding the $3.7b versus $21b figures.

Page v of the report describes the financial costs of obesity as $3.7b, and the non-financial costs as $17.2n, totalling $21b. That’s probably where the two figures came from.

the fat tax can be structured such that the tax payable is directly proportional to the health-related factors in the food, then it will be a specific well-targeted instrument.

There’s an assumption here that a health-related factors are linear measures found in foods. For an example such as dietary kilojoules, this may be a fair assumption. However, if you are trying to promote a balanced diet, avoiding excesses in all directions (including too much cabbage) this may be more difficult.

That’s funny, I thought it is exactly because of the inelastic demand that we have an excise on alcohol and cigarettes…

I was surprised by the claim that alcohol and cigarettes are price-inelastic. (The same claim appears in the report.) I don’t know the reality, but I would have thought that there’s a decent amount of elasticity. Was the intention of the cigarette and beer excises purely to raise revenue? Seems a tad cynical to me.

That’s because there is no evidence that the GST concession has resulted in cheaper prices for healthier food, see above.

You lost me here. Can you explain why you say there is no evidence?

So you’re saying that it won’t convince people to skip meals or otherwise change their overall eating patterns? Well, duh. “Hmm, I can’t afford that $20 Big Mac, but I’d rather go hungry than eat this $5 tuna sandwich instead!”

Consider the statement “The price of a Big Mac is now $20, and I must eat them, so I can no longer afford gym membership and school books.” I am not saying this is rational or realistic, but I think you have to debunk that people will think like this in order to properly dismiss the author’s claims about inelasticity.

In my world, the junk food excise would be used to reduce the price of non-junk food by a corresponding amount.

Equivalently, the money raised by the fat tax could go to reducing income tax and increasing welfare.

I saw some other strawmen in the report:

if a tax is effective in switching consumption away from a product, it will not raise much revenue

The try to make that sound like a bad thing. The goal of the tax is to change behaviour, not to raise revenue. Low revenues would count as a huge success (because of the $21b in indirect savings!)

If a fat excise, sugar excise and salt excise were all introduced then there would be up to 15 different tax rates potentially applicable to any food item.

Where did the number 15 come from? If 15 is too high, it could instead be implemented with five or three or two or whatever figure was electorally acceptable.

Unlike Alastair, I am not yet swayed enough to say “Let’s do it!” But not for any of the reasons outlined in this report.

There’s an assumption here that a health-related factors are linear measures found in foods. For an example such as dietary kilojoules, this may be a fair assumption. However, if you are trying to promote a balanced diet, avoiding excesses in all directions (including too much cabbage) this may be more difficult.

Agreed, but an assumption of linearity is a pretty good approximation IMHO. Of course it remains to be seen, but I would think that it is very possible that a tax could be constructed that reflects the overall nutritional value of food, if even on a very rough order of approximation.

I was surprised by the claim that alcohol and cigarettes are price-inelastic. (The same claim appears in the report.) I don’t know the reality, but I would have thought that there’s a decent amount of elasticity. Was the intention of the cigarette and beer excises purely to raise revenue? Seems a tad cynical to me.

There is inelastic demand for cigarettes simply because of the addictive nature of tobacco. If your body is telling you that you need to smoke no matter what the price, I understand that it is very difficult to tell it otherwise.

I guess it could be seen as cynical to impose the excise simply because of the inelastic demand for these items. However unlike junk food it is a lot more difficult to find low-risk alternatives to cigarettes and alcohol. So the emphasis more naturally falls on mitigation rather than prevention. The cost of these measures has to be paid for somehow, and I had always assumed assumed that this was the real reason for imposing excise on cigarettes and alcohol. So whether or not it is cynical depends how the money is used, I guess.

That’s because there is no evidence that the GST concession has resulted in cheaper prices for healthier food, see above.

You lost me here. Can you explain why you say there is no evidence?

I had made the point earlier that junk food being more expensive than non-junk. There is no evidence that the GST has resulted in cheaper prices for healthier foods because there aren’t cheaper prices for healthier foods.

It’s easy to see why though: the GST simply wasn’t targeted at promoting health. We can not expect the GST to solve our health problems any more than industrial relations reform or sending troops to Iraq. It’s a red herring.

Consider the statement “The price of a Big Mac is now $20, and I must eat them, so I can no longer afford gym membership and school books.” I am not saying this is rational or realistic, but I think you have to debunk that people will think like this in order to properly dismiss the author’s claims about inelasticity.

I thought I had done so by pointing out that even though the overall demand for food in inelastic, this does not imply that the demand for different types of food is also inelastic. This does not follow and the burden of proof is on the author.

I strongly suspect that at any given price point junk food can easily be substituted with healthy food and the same level of demand will result, all else being equal. Our demand for food is is based upon our underlying biology and this, as with alcohol addiction, is the source of the price inelasticity.