By simultaneously offering to reduce weight gain and boosting government revenue at a time when many states are cash-strapped, soda tax proposals have leapt from the pages of academic journals into the legislative agendas of statehouses. Indeed, soda taxes seem to be a win-win for everyone but drink manufacturers and individuals clinging to their soft drinks.
Against this are the traditional arguments against sin taxes: they require people to keep on sinning or the tax revenue will shrink, and they tend to affect poorer people disproportionately, which is to say, they are regressive. The question this analysis set out to answer is whether, based on the current evidence, soda taxes would have the public health benefits their advocates claim – namely, whether they will result in lower calorie consumption or weight loss. Our analysis suggests there is reason to be skeptical of such claims.
Advocates for soda taxes have selectively interpreted the available data on SSB consumption and weight gain to build tax models which magnify the certainty of a link. Whenever complexity intrudes on this model – that consumption patterns vary dramatically, that high soda consumption is confounded by other high calorie food consumption and low energy expenditure, that substitution effect modeling suggests, at best, minimal impact on weight loss – this complexity is ignored.
Worse, the reaction of some scholars to these challenges is to attribute any objections to the corrupting influence of industry funding. For example, in their perspective piece for the New England Journal of Medicine – “Ounces of Prevention” – Brownell and Frieden note that “studies that do not support a relationship between consumption of sugared beverages and health outcomes tend to be conducted by authors supporting the beverage industry.” Yet their footnote points the reader to just one of such study, the aforementioned Forshee et al. As Gibson notes in her systematic review, there are comparatively few industry funded studies in this area. There are also non-industry supported analyses doubtful of the link, such as Bachmann et al. and Wolff et al.
Ironically, there is evidence of publication bias in the other direction, along with evidence for research findings being mischaracterized in the direction of establishing a link between weight gain and SSBs as opposed to other sources of calories. When scholars assess the evidence outside the context of soda tax advocacy – such as the USDA’s advisory committee discussion on dietary recommendations – their assessment of the link between soda and weight gain is much more qualified, their criticism of the quality of the evidence for such a link much harsher.
So where does this leave us?
Guzzle soda without cutting back on the rest of your diet or exercising more, and you will gain weight. But while some people believe soda calories are the equivalent of super calories when it comes to putting on weight, the evidence - as the advisory committee for the US Department of Agriculture notes - is that soda calories are the same as calories from other food.
The best available evidence shows that those children and adults who are overweight and who are high consumers of sugar sweetened beverages would benefit most from reducing consumption. Would a soda tax achieve a meaningful reduction for this group? Given the hypothesized calorie reductions from the tax proposals discussed here, and all the caveats about substitution effects and confounders, it’s hard to conclude anything more than that there might be modest benefits for some consumers.
At the same time, it is also possible that the people who consume the most soda are the most price insensitive in which case, the impact of a soda tax on those whom it is directed at the most will be limited. Add to this the conclusion by Finkelstein et al. that low income consumers – the constituency which needs to reduce calorie intake the most – would probably find ways to bypass any additional soda taxation, and there are further grounds for skepticism about the public health payoff from increasing soda taxes.
In sum, if the key weakness in the nutrition research is the degree to which soda consumption leads to weight gain independent of any other behavior, the key weakness in the economic literature is the wide range of compensated price elasticities. Research provides numbers to give you whatever outcome you want, but little to establish what the right number actually is. In essence, a soda tax is fundamentally experimental: We will only know which of the various competing models and interpretations and arguments are right if the tax is actually imposed.
One other argument for soda taxes is that they would encourage people to switch to low or zero calorie drinks, which could lead to a substantial reduction in calorie intake. Again, there is no guarantee that this would happen given the range of possible compensated price elasticities – and the fact that the relative difference in taste between diet and regular drinks may lead consumers to reject low calorie drinks as substitute goods.
There has been a very modest trend toward increased consumption of diet drinks at the expense of sugar-sweetened drinks from 2005 (29.5 percent of all such drinks) to 2010 (29.9 percent), according to the trade journal Beverage Digest. But the real change may come in the next two years, says editor and publisher John Sicher, as companies roll out new naturally and artificially sweetened low calorie drinks with better taste profiles.
In other words, the market may provide a solution that will be more effective than taxes in convincing people to reduce their SSB intake, by providing compelling, lower calorie substitute goods. There is no economic model that can predict whether this will be the case, but the current trend is at least in the right right direction. Unfortunately, the evidence that increased taxes will have a meaningful impact on America’s growing obesity epidemic is weak. The only real certainty is that a soda tax will generate revenue, leaving all soda consumers, overweight or not, to swallow the bill.
Back to the homepage
The Statistical Assessment Service (STATS)
933 N. Kenmore St., Suite 405
Arlington, VA 22201