My thanks to Baylen Linnekin, Russell Saunders and Jennifer Harris for their thoughtful responses to my original article. As Linnekin largely agrees with me, this post will mainly address the arguments of Saunders and Harris.
Firstly, I should say that I do not hold some of the opinions that Harris and Saunders consider to be truisms. Saunders says ”that obesity is a major U.S. public health problem is not a subject of much dispute” while Harris says that “All agree that parents cannot solve the childhood obesity crisis on their own.” In fact, I do dispute that obesity is a “public health problem.” I don’t share the currently fashionable view that a public health problem is merely the aggregate of a nation’s private health problems. Obesity differs from genuine public health issues, such as unclean drinking water, pollution, and tuberculosis, in that it is not infectious and it can be prevented and cured without government intervention. It a private health problem – or, more correctly, it is a risk factor for private health problems.
I am loath to start a sentence with the words “as a parent,” but as a parent I reject Harris’s assertion that “parents can’t compete with the overwhelmingly unhealthy food environment surrounding their children as soon as they step outside the front door.” Parents can prevent their children becoming obese, particularly when they are young. Collectively, therefore, parents can “solve the childhood obesity crisis,” if it must be put in those terms. Similarly, parents can prevent themselves from becoming obese. The evidence is all around us. Even in the United States, with its supposedly “obesogenic” environment, two-thirds of adults and 83 percent of children are not obese. Obesity is not rare enough to be called deviant, but nor is it normal enough to be viewed as an unavoidable consequence of forces that are beyond the individual’s control.
In a free society, the question of whether government should introduce legislation to tackle obesity does not depend on obesity being a “pubic health problem” or even a “crisis.” Intervention can only be justified if obesity results from market failure or creates negative externalities for those who are not obese. Harris suggests that both criteria have been met. She uses what is essentially a market failure argument when she talks about the need for “widespread systemic changes to the food environment.” The implication is that consumers are denied real choice because so many products contain sugar, salt, and fat. This is the famous “obesogenic environment” in which healthy choices are supposedly difficult, if not impossible, and consumers are unable to satisfy their latent desire to eat cabbage and broccoli.
But look at the choices on the shelves, even in inner cities. There have never been so many low-fat, low-sugar, and low-calorie options in stores and supermarkets, never so many fruits and vegetables from around the world. You need look no further than the soda market to see the range of options available to the weight-conscious consumer. The Coca-Cola company launched its first low sugar brand (Tab) in the 1960s, and it now produces Diet Coke, Coke Zero, and Coke Life. Diet Coke and Coke Zero contain no sugar, and Coke Life is a low calorie version of Coke, made with a combination of sugar and sweeteners. Nearly all soft drink companies produce similar low-calorie, zero-calorie, and sugar-free varieties. All are widely advertised and all are available on the same shelves, in the same stores, and for the same price as their more sugary cousins. It is very difficult to argue that consumers are nudged, let alone coerced, into buying the high-calorie variants. If they buy them it is because they want them.
As for negative externalities, government could perhaps justify action if obesity was associated with costs to taxpayers. Harris mentions $190 billion of obesity-related health costs in the United States, some of which comes from the general taxpayer. These are not net costs however. Obesity, like smoking, is associated with savings to the taxpayer as a result of less government expenditure for welfare in old age, such as pensions. I am not aware of any relevant studies using American data, but research from the Netherlands found that the lifetime healthcare costs of obese people were significantly lower than those of healthy weight. If we look at government expenditure in the round, it seems unlikely that the state would spend less money – or that it would be a “boon to our pocketbooks,” as Saunders puts it – if obesity was eradicated.
It is, then, far from obvious that obesity creates negative externalities deserving of government action. Since obesity is neither contagious nor the result of factors that are beyond the individual’s control, it is not a public health problem in the traditional – and, I would argue, meaningful – sense. It is merely a risk factor for private health problems, such as diabetes, which might result from it. Sugar fits into the picture only insofar as it is a source of calories, that is, insofar as it is food. In other words, it is a risk factor for a risk factor. Since there is a wide range of choice and little evidence of market failure, it is fair to assume that consumers’ revealed preference for high calorie food is their true preference. Any attempt to make their preferred choices more difficult will result in a welfare loss. Not everybody sees pristine health and maximum longevity as their goal. When I buy a Coke, I am aware that it provides “zero nutritional value,” as Harris puts it. I don’t buy it for the nutrients. I buy it because I like it.
But let us say that there was a market failure to be corrected and that obesity was a burden on taxpayers. Would soda taxes be a sound response? I say no. As I said in the initial post, the impact of sin taxes on food and drink in the real world has been trivial at best. The costs have inevitably exceeded the benefits (not least because there have been no benefits), and the burden on taxpayers has not shrunk, it has grown. Saunders acknowledges that every soda tax yet tried has been “too low to effect much change in obesity rates.” He also recognizes that local soda taxes lead to cross-border sales and therefore suggests that “a broader geographic scope might well enhance efficacy.” I assume from this that his solution is to have high soda taxes at the state or federal level. Fine. No one denies that such taxes could have an impact on consumption – and, perhaps, obesity – if they were high enough. The question is whether the negative consequences would outweigh the positive consequences. The shipwreck of the Danish fat tax should be our landmark here: Denmark’s tax witnessed cross-border sales, inflation, substitution effects, job losses, political unpopularity, and a disproportionate impact on the poor that ultimately led to repeal. Such negative consequences are likely to grow with the size of the tax. Ultimately, if taxes were raised to the eye-watering levels that would be required for them to have a measurable impact on health, consumers would be served by the black market, with all the negative consequences that it entails.
As for the positive consequences, do not underestimate substitution effects. This is not tobacco. Quitting is not an option. We have to eat and drink something, and for sound evolutionary reasons we are drawn towards food and drink that tastes nice (that is, food that has lots of calories). Saunders says: “Snowdon suggests that consumers will simply switch to an alternate form of liquid calories. Pepsi drinkers will just become drinkers of beer, milkshakes, or juice. But this argument is belied by actual experience.” Actually, that is precisely what Fletcher et al. found in their study which concluded that “reduction in soda consumption is completely offset by increases in consumption of other high-calorie drinks.” And, as the Danish experience showed, there is ample scope to maintain calorie consumption by shopping in cheaper stores and buying cheaper brands.
On the issue of cost-effectiveness, Saunders criticizes my cost-benefit analysis of what a soda tax would mean for Londoners. In fact, my methodology is no different from that of the Children’s Food Campaign (CFC), whose figures I analyze, and my estimate of the cost to Londoners (£2.6 billion) could be decimated, halved, or quartered and still be much larger than the claimed savings (£39 million). Perhaps I can make things clearer by using figures for the whole United Kingdom. It is said that obesity costs the British health service £5 billion (this doesn’t include savings; the net figure would be lower, possibly even negative). The CFC’s soda tax would, by their own estimate, cost £1 billion. A 20 per cent tax on sugary drinks would therefore have to reduce obesity-related healthcare costs by more than a fifth for it to reduce the overall tax burden. This is highly unlikely and the CFC’s estimates for London indicate that the savings would indeed be much lower than this. Russell’s comment that “savings from a soda tax” may not be “as clear-cut as advocates claim” is a huge understatement. The inelasticity of demand for sugary drinks means that savings are quite unattainable.
The best argument for soda taxes is, as Saunders notes, that they might shift more of the obesity-related tax burden from the slim to the obese. But taxes on sugary drinks are a very clumsy way of doing this; lots of obese people don’t drink them, and lots of slim people do. I would find this argument more convincing if soda tax campaigners outlined which taxes were going to be cut in order for the overall tax burden to fall.
Saunders is right to say that soda tax revenue need not inevitably be absorbed by general expenditure and new bureaucracies, but past experience creates a strong presumption. Denmark was exceptional in this respect by lowering its income tax by the same amount as the fat tax was expected to raise (and increasing its income tax when the fat tax was abandoned). It is difficult to imagine many governments being prepared to bring in unpopular taxes on food and drink unless they were going to bring in more revenue. Politicians would do better to follow Linnekin’s advice and stop subsidizing sugar and corn syrup – government interventions which really are a drain on the taxpayer.
For the sake of argument, let us say that a soda tax could shift the relevant healthcare costs efficiently to the obese on a “polluter pays” principle. Saunders acknowledges that this would place a greater financial burden on poor people – including, I might add, poor people who are not obese. I note that the study Saunders cites as evidence that “low-income soda drinkers are much more likely to be obese” found only a modest association between soda consumption and weight gain amongst women when other factors, such as watching television and physical inactivity, were controlled. Amongst men there was no association at all. This should serve as a reminder that the evidence linking sugar and soda to weight gain is, as a 2007 review noted, “inconsistent.” In my view, the evidence is nowhere near strong enough to justify making soda drinkers the scapegoat for the obesity “crisis” while letting couch potatoes off the hook (physical inactivity hovers over this whole issue like Banquo’s ghost; despite being at least half the problem, it cannot be taxed or banned and therefore does not get the attention it deserves from the public health lobby). Soda taxes would be discriminatory and unfair even if they did not take a disproportionate sum of money from the poor, but their invariably regressive nature makes them even more objectionable. Taxing food and drink is unlikely to make the poor slim, but it will certainly ensure they stay poor.
Saunders and Harris both acknowledge that a soda tax would have a limited effect “as a stand-alone method.” Harris proposes “a combination of taxes, limits on sales and marketing to young people, increased understanding of the long-term negative consequences of poor diet in childhood, and changes in norms around consumption of unhealthy foods and beverages will all be required.” This, as she explicitly states, is the application of the tobacco control model to diet.
I am almost – though not quite – tired of reminding people that anti-smoking campaigners spent decades denying that such a “slippery slope” exists. In 1994, RJ Reynolds was mocked for producing an advert which asked: “Today it’s cigarettes. Will high-fat foods be next?” They were only wrong about the ingredient that would come under fire, although given how rapidly the scientific consensus switched from fat to sugar, that is perhaps understandable. Reynolds asked the question because they knew that the American citizen of the 1990s would find the idea of waging war on “Big Food” to be self-evidently absurd. Since that assumption can no longer be taken for granted, we need to ask whether tobacco-style regulation is an appropriate and effective approach to diet.
Again, I say no. Food is not tobacco. Obesity is not lung cancer. And, contrary to some hyperbolic claims made in recent years, sugar is not addictive. As discussed above, taxes are not as effective when there are a range of substitute products available. As for advertising bans, which Harris dwells on, the economic evidence is very clear that advertising can stimulate demand for a particular brand but has little or no effect on demand for established product categories, such as soda or hamburgers, as a whole. I appreciate that this goes against some deeply held beliefs about the supposedly coercive nature of advertising, and I know that there are public health academics who insist that primary demand is driven by marketing, but I find the evidence of economists who have looked at numerous different products in dozens of different countries to be more compelling.
If there is a lesson to be learned from tobacco control, it is that it never ends. Harris and Saunders make it clear that soda taxes are just the start. What, then, is the endgame? How much obesity would they tolerate? How much tax would be too much? If we are talking about a whole suite of policies, the public has a right to know what they are signing up to before taking the first step. It is no coincidence that the anti-smoking movement began with sin taxes and advertising bans. Does the anti-obesity lobby also favor the more aggressive policies that have turned the anti-smoking crusade into a war of attrition? Can we expect soda drinkers, like smokers, to be treated like second -class citizens? Can we expect the demonization of the soda industry, which has already begun in California? Lawsuits? Point of sale bans? Mandatory product reformulation? Bans on certain flavors? Bans on drinking Coke in public places?
These are not frivolous questions. I maintain that every food and drink tax yet tried has been ineffective. Harris and Saunders argue that higher taxes would work better – in extremis they might be right – but say that they would still have a limited effect unless combined with a range of measures lifted from the tobacco control playbook. Soda taxes cannot, therefore, be judged in isolation. Unless we are told what will accompany them, we are being sold a pig in a poke.