Should we tax soft drinks? A look at the data…

Proposals to tax unhealthy food, and possibly subsidize healthier alternatives, have been a topic of significant controversy over the last few years. The idea behind these proposals is to discourage the consumption of food that leads to obesity, heart disease, and diabetes by financially penalizing their consumption, just like the highly-successful tobacco taxes of the past. In today’s blog post, we look at the data behind the most prominent example of this form of legislation: the proposal to tax sugar-sweetened beverages.

Sugar-sweetened beverages such as soft drinks appear to be the largest statistical driver of the obesity epidemic in the United States. So it’s no surprise that some jurisdictions have proposed taxing soft drinks to increase their price and thereby discourage their consumption. If the most ardent capitalists would object to restricting individual choice through taxation, perhaps they should look back to The Wealth of Nations, in which Adam Smith himself suggested that “sugar, rum and tobacco are commodities which are nowhere necessaries of life, which are become objects of almost universal consumption, and which are therefore extremely proper subjects of taxation.”

Regardless of what Smith says, there remain common arguments against soft drink taxes:

(1) Won’t soft drink taxes be “regressive”? Regressive taxes are those that disproportionately affect the poor, who typically spend a larger proportion of their income on food than the wealthy. In fact, research into the impact of taxes on soft drinks reveals that such taxes have the potential to be most beneficial to low-income people, particularly because obesity is a “regressive disease”–disproportionately affecting the poor and minorities. The poor tend to consume more soft drinks, and are more sensitive to higher prices than the wealthy; higher taxes on soda, in turn, seem to direct them towards a great, free alternative source of refreshment: water. Soft drinks, after all, are not a necessary part of the diet, and don’t contribute to nutrition.

(2) Aren’t people who consume soft drinks are making the conscious choice to risk personal obesity? In fact, recent survey-based research suggests that people are rarely aware of what the risks of their food choices are, even after education campaigns, because so many unhealthy foods are disguised behind healthy-appearing labels, and because marketing is so effective. Few people realize, for example, than a 20 ounce bottle of Coca-cola contains over 15 teaspoons of sugar and 240 calories; most people grossly underestimate the number of calories they eat, even if they’re professional nutritionists. The result is “information asymmetry”, in which the two parties in the transaction–consumers and manufacturers–have vastly different knowledge about what’s contained in the product being sold. People also have “time-inconsistent preferences”, meaning that we (and children in particular) tend to privilege short-term gratification over long-term risk, often to our detriment. Meanwhile, about $1.62 billion has been spent to market soft drinks and related foods, about $870 million of which is directed towards children under 12 who are often capable of purchasing these foods without parental supervision, at school.

(3) Shouldn’t the government stay out of matters of personal choice? Well, it turns out that the government–that is, the taxpayers–are paying for the problem already. The personal choices of millions of Americans are not just personal–they carry “externalities” out to the rest of the population. The medical costs of obesity are about $147 billion per year, half of it paid by Medicaid and Medicare. But our government is also spending on programs that are actively promoting obesity, particularly agricultural subsidies to large conglomerate corn producing companies for the production of high-fructose corn syrup. So the government is already deeply involved in deciding the marketplace of choices, and sets nutritional standards for school meals. Besides, the public function of government does include preserving public health by setting minimum standards for safety, such as requiring seat belts.

Is there enough data to support limiting soft drink intake?

In the mid-1990s, children’s intake of sugar-sweetened beverages in the United States surpassed their intake of milk (see the graph below). In conjunction, the daily calorie intake of the American child rose, and diabetes diagnoses skyrocketed. Statistically speaking, every extra can of sugar-sweetened beverage consumed by an American child increases her chances of becoming obese by 60%.

A series of studies have suggested why sugar-sweetened beverages in particular may have an unusual predilection for inducing obesity and diabetes, and why limiting their intake may be highly beneficial to the overall population. Sugar-sweetened beverages appear to have particularly low satiating properties, meaning that they deliver calories without delivering a sensation of being full. Intake of solid foods don’t seem to differ between people who drink soda versus water during meals, meaning that those drinking soda don’t feel the calories in their soda and don’t substitute them for less solid food; the result is about 200 kilocalories per day more consumption among those drinking soda during controlled experiments. The “addicting” properties of these high-sugar drinks seem to discourage their discontinuation once people start consuming them.

Furthermore, sugar-sweetened beverages have a high “glycemic load”, delivering a large amount of sugar per unit food, and as a result it’s not surprising that large, well controlled studies show a significant effect of soda consumption not only on obesity but also on diabetes. Similar evidence shows a higher risk of heart disease among those with high soda consumption after correcting for potential confounding variables, likely because of the contribution of refined carbohydrates to elevating triglyceride levels and blood pressure. Conversely, reducing soft drink consumption has been shown to improve obesity and related disease correlates. One meta-analysis sponsored by the beverage industry has been famously quoted for having disputed these other studies; it incorrectly gave large weight to several small studies, and when re-analyzed by an independent group, its results supported the same soft drink-body weight relationship as the other studies.

But wouldn’t simply educating people about soft drinks be sufficient to reduce their consumption? Unfortunately, the population-level efficacy of education-only interventions remains markedly poor, in comparison to marketplace-altering alternatives that have been the only effective interventions to have reduced obesity at a population level.

As compared to other foods, there are two factors related to soft drinks that make them particularly appropriate for targeting through taxation: there are just a few large manufacturers of these beverages who are doing quite well economically, making them relatively easy to regulate and of little economic detriment to add small taxes to; and, more importantly, soft drinks are the single largest contributor of calorie intake in the United States diet.

Will soft drink taxes actually work?

Two issues come into play when calculating what the population-level impact of a soft drink tax might be: elasticity and cost-shifting. Price elasticity refers to the degree to which a consumer will alter their purchasing behavior of a product in response to a change in price. If a good is very inelastic, often because there’s no alternative or because the product is a basic necessity, then even a marked increase in price won’t change how much people buy a product. In the food realm, eggs are a relatively inelastic product that people buy even after significant price hikes (mmm, omelets). But soft drink consumption is rather elastic, such that a 10% increase in price amounts to about a 8% drop in consumption, conservatively; some reports from Coca-cola and Pepsi actually report an even higher effect of price increases. Of course, the degree of elasticity can change with the size of a tax; very small taxes have not shown great effect in some settings, but even small taxes on beverages in some states show significant impact on populations that are most at risk for obesity from sugar-sweetened beverages: minorities, low-income people, and kids who purchase soft drinks at school.

The effect of a tax is also contingent on how much manufacturers actually shift taxes placed on them to the consumer. “Under-shifting” occurs in some markets where the manufacturer tries to absorb a tax and keep the consumer price of the good nearly the same. It turns out that the beverage industry is classically an “over-shifter“; like airlines, the beverage industry has viewed taxes as an opportunity to collectively raise prices throughout an oligopolistic market, such that consumers have no cheap alternatives. This means that taxes would likely result in significant price increases–and subsequently notable declines in consumer demand due to the elasticity of soft drink sales. The result is a large decline in calories consumed, but a lower tax revenue over time because over-shifting implies less quantity of products produced and therefore less revenue after a while.

What about unanticipated side-effects?

Critics like to dismiss public health advocates as crazy people trying to experiment with the public (because replacing chemical drinks with water seems very dangerous); the idea that public health programs would have vast unanticipated consequences is a favorite criticism that is too vague to effectively respond to. But when it comes to taxation, we can anticipate what some of the public health effects of regulation might be.

First, the consequences of any given tax proposal depends on what kind of tax is instituted. Two major types of tax are under consideration for sugar-sweetened beverages: excise taxes and sales taxes. Excise taxes are fees per ounce of a beverage produced. The fee is placed on the manufacturer, who then passes it to the consumer. This is easier to collect than a sales tax, and a more reliable type of tax from a public health perspective, since major manufacturers would be taxed (easy to identify who needs to pay, and prevent rampant corporate tax evasion) and even if the industry tries to reduce prices to the consumer, the incentive to the manufacturer of producing less sugar-sweetened beverages and more alternatives like bottled water or low-sugar drinks remains the same. An excise tax would also produce predictable revenues based on manufacturing data, allowing government agencies to budget ahead of time for programs that could be funded by the tax (more on that below) and could be indexed for inflation to avoid diminishing effect over time.

In contrast, a sales tax on sugar-sweetened beverages, which would charge a customer a fee based on the product’s price, is likely to be more problematic. Consumers may be encouraged to buy in bulk because the cost per unit volume would be lower. Retailers would face the burden of executing the tax, and revenue agencies would have trouble ensuring that it was paid for by everyone from big-box stores to local gas stations. Inconveniencing restaurants and retailers has also classically resulted in great political opposition, as opposed to just inconveniencing beverage manufacturers.

One key ingredient of any tax proposal would be that the tax should not just apply to “soda”. Over the past several years, a myth has propagated that because diabetes continues to increase while soda sales have decreased overall, soda cannot possibly be contributing to the diabetes epidemic. This ignores the time-delay between consumption and disease, but it also ignores the fact that soda consumption has been substituted for with other high-sugar drinks like ready-to-drink coffee and tea, as well as a variety of fruit and sports drinks, most of which have high sugar calorie content. So a soft drink tax on “sugar-sweetened beverages” should not just apply to cola. If it applies to all sugar-sweetened beverages based on the amount of sugar per ounce, for example, it’s also unlikely to result is a significant “substitution effect” to other types of food. If diet beverages are made exempt from the tax, it may encourage consumers to switch to healthier alternatives to sugar-sweetened drinks, though this may diminish revenue.

Research has nevertheless shown that, just as people do not recognize their calorie load from soda and still eat just as much food on the side, similarly they do not miss the calories from soda and substitute the decreased soda for other unhealthy foods. So, ultimately, the side-effects of taxation are critically dependent on what kind of tax is instituted, but the data available suggest that an excise tax would be unlikely to result in massive substitution towards other high-calorie products or generate public health harm.

How much can taxes do?

By using a conservative estimate of ongoing demand for soft drinks, multiplied by the price elasticity of soda consumption, and taking into account the available data on cost shifting in the beverage industry, economists have been able to project that a penny-per-ounce excise tax on sugar-sweetened beverages would likely reduce the consumption of these beverages by 24% (about 50 kilocalories per day per person, or 5 lbs of weight loss per year). Such levels of weight loss are associated with significantly decreased risk of heart disease and diabetes.

Given the one-cent-per-ounce tax and expected changes in demand, such a tax would also produce $79 billion over the first five years of its implementation, and $13 to $15 billion in just its first year. The first year estimate is thirty times the amount that the nation’s largest funder of childhood obesity prevention spent in five years.

To determine what any given state or major city could earn in revenue, the Rudd Center for Food Policy and Obesity (the nation’s leader of research on the soft drink tax subject) has put together an online tax revenue calculator that we can use to estimate the impact of the tax in our locale:

What could we do with this money? Several proposals have been put forward, but the most common include subsidizing fresh fruits and vegetables, funding better food in schools, and addressing aspects of the built environment in cities that contribute to sedentary behavior and obesity (e.g., park space).

Will taxes pass legislatures?

So far, efforts by cities like San Francisco and New York, as well as hearings on the proposal by the US Senate, have been rebuffed by a massive industry campaign against soft drink taxes. Direct threats from Coca-cola and related manufacturers have been made on school administration officials. The White House has distanced itself from the soda tax proposal and pursued more voluntary nutrition labels from food manufacturers, although the quality of such initiatives is seriously in question. Meanwhile, the beverage industry has created “Americans Against Food Taxes” and hired Obama’s former communications director to fight reforms such as the soft drink tax. A recent review reveals a full list of the industry’s lawfirms and lobbyists who are mobilizing against the idea.

In contrast, members of the general public appear supportive of the idea. Fifty-two percent of New Yorkers supported the tax, and the number went up to 72% when respondents were told the revenue would go towards obesity prevention. The issue also has increasingly entered into the public eye as advocates of the tax, such as the Rudd Center’s Kelly Brownell, have appeared on major media outlets describing the initiative. Brownell has been spearheading a number of public education measures on the topic (see the YouTube video below). Whether the taxation will follow the line of tobacco taxes and be implemented to public health benefit, or whether it will be another lost opportunity for revenue during recessions, remains to be seen. But if “big food” acts like “big tobacco”, the battle over the regulatory rights of government could go on for some time.