|Home | About | Journals | Submit | Contact Us | Français|
Consumption of sugar-sweetened beverages (SSBs) has been linked to risks of obesity, diabetes, and heart disease,1–3 making a compelling case for reduced consumption. SSBs are beverages which contain added naturally-derived caloric sweeteners such as sucrose (table sugar), high-fructose corn syrup, or fruit juice concentrates, all of which have similar metabolic effects.
Taxation has been proposed as a means for reducing intake and thereby lowering health care costs, as well as for generating revenue that governments can use for health programs.4–7 Presently, 33 states have sales taxes on soft drinks (mean = 5.2%), but the taxes are too small to affect consumption and the revenues are not earmarked for programs related to health. This paper examines trends in consumption of SSBs, evidence linking these beverages to adverse health outcomes, and approaches to designing a tax system that could improve nutrition and help the nation recover health care costs associated with SSB consumption.
In recent decades, SSB intake has increased around the globe, for example doubling between 1999 and 2006 in Mexico in all age groups.8 Between 1977 and 2002, per capita caloric beverage intake doubled in the U.S. across all age groups9 (Figure 1). The most recent data (2005–2006) show that U.S. children and adults consume about 172 and 175 kcal per capita daily from sugar-sweetened beverages, respectively.9
The relation between consumption of SSBs and body weight has been examined in many cross-sectional and longitudinal studies and has been summarized in systematic reviews.1, 2 A meta-analysis found positive associations between SSB intake and body weight that were stronger in longitudinal studies than in cross-sectional studies and in studies not funded by the beverage industry than in those that were.2 A meta-analysis of studies in children supported by the beverage industry10 was interpreted as showing no evidence for an association between SSBs and body weight, but it erroneously gave large weight to several small negative studies; when correct weighting was used, the meta-analysis summary supported a positive association.11 A prospective study among middle school students over two academic years found that the risk of becoming obese increased by 60% for every additional serving of SSB per day.12 In an 8 year prospective study of adult women, those who increased SSB consumption at year 4 and maintained this increase gained 8 kg, whereas those who decreased SSB intake at year 4 and maintained this decrease gained only 2.8 kg.13
Short-term clinical trials provide an experimental basis for understanding how SSBs might affect adiposity. Tordoff and Alleva14 found that total energy intake and body weight increased in subjects given 530 kcal per day of SSBs for three weeks, but decreased when subjects were given non-caloric sweetened beverages for the same time, relative to a period of no beverage provision. Raben et al.15 reported that obese subjects gained weight when given sucrose primarily in the form of SSB for 10 weeks, whereas they lost weight when given non-caloric sweeteners for the same length of time.
Four long-term randomized controlled trials of SSBs and body weight have been reported, and showed the strongest effects in overweight individuals. A school-based intervention to reduce carbonated beverages was conducted with 644 students ages 7–11 years in the UK using a cluster design.16 After one year, the intervention group showed a non-significantly lower mean BMI compared to controls and a significant 7.7% lower incidence of obesity. In a study with 1140 Brazilian school children ages 9–12 designed to discourage SSB consumption, no overall effect on BMI was observed during the 9-month academic year.17 Among students overweight at baseline, BMI decreased non-significantly in the intervention group vs. control; among overweight girls, the difference was statistically significant. In another clinical trial, 103 Boston high school students were assigned to a control group or an intervention group that received home delivery of non-caloric beverages for 25 weeks. BMI was non-significantly reduced in the overall intervention group, but among individuals in the upper baseline BMI tertile, the intervention produced a significant decrease in BMI (−0.63 vs. 0.12 kg/m2).18 The effects of replacing SSBs with milk products were examined among 98 overweight Chilean children.19 After 16 weeks, the percentage body fat increased non-significantly less in the intervention group (0.36 vs. 0.78%), whereas accretion of lean mass was significantly greater in this group (0.92 vs. 0.62 kg).
Three prospective observational studies conducted in US Nurses, Finnish men and women, and Black women each found positive associations between SSB consumption and risk for type 2 diabetes.13, 20, 21 Among the 91,249 women in the Nurses’ Health Study II followed for eight years, the risk of diabetes was nearly doubled for women consuming one or more servings of SSB per day13; about half of the excess risk was accounted for by greater body weight. Among Black women excess weight accounted for most of the excess risk.
Among 88,520 women in the Nurses’ Health Study, risk for coronary heart disease increased by 23% for women who consumed one SSB serving per day and 35% for two servings or more.3 Increased body weight explained some but not all of this association.
A variety of behavioral and biologic mechanisms may be responsible for the associations between SSB consumption and adverse health outcomes, with some links (e.g., SSB intake and weight gain) better established than others. The well-documented adverse physiological and metabolic consequences of high intake of refined carbohydrates such as sugar include elevation of triglycerides and blood pressure and lowering of HDL cholesterol, which would be expected to increase risks of coronary heart disease.22 SSBs, due to their high glycemic load, would be expected to increase risk for diabetes by causing insulin resistance and also through direct effects on pancreatic islet cells.23 Observational research found that consumption of SSBs, but not non-calorically sweetened beverages, is associated with markers of insulin resistance.24
SSB intake may cause excessive weight gain due in part to the apparently poor satiating properties of sugar in liquid form. Indeed, calorie compensation for energy consumed as a beverage compared to solid food is less complete at subsequent meals.25 For example, in a study using 7-day food diaries from 323 adults, energy from beverages added to total energy intake and did not appear to displace other sources of calories.26 Consistent with the adult data, a study of school-age children indicated that those who drank 9 oz or more of SSB per day, on average, consumed nearly 200 kcal per day more than non-drinkers.27
Short-term feeding studies support this mechanism. Among 33 adults given identical test lunches on six occasions with varying beverage type (sugar sweetened cola, non-caloric cola or water) and amount (12 vs. 18 oz),28 solid food intake did not differ across conditions, resulting in significantly greater total energy consumption when the SSB was served.
SSBs may also affect body weight through other behavioral mechanisms. Whereas solid food intake is characteristically coupled to hunger, individuals may consume SSBs to satisfy thirst or for social reasons in the absence of hunger. SSBs may also have chronic adverse effects on taste preferences and food acceptance. Individuals, especially children, who habitually consume SSBs rather than water may find more satiating but less sweet foods (e.g., vegetables, legumes and fruits) unappealing or unpalatable, resulting in poor diet quality.
Economists agree that government intervention in a market is warranted when there are ‘market failures’ that result in less than optimal production and consumption.29, 30 Several market failures exist for sugar-sweetened beverages. First, many persons make consumption decisions with imperfect information, failing to fully appreciate the links between consumption and health consequences. These decisions are likely to be further distorted by the extensive marketing that can raise the perceived benefits of consumption. A second failure results from time-inconsistent preferences (decisions that provide short-term gratification but long-term harm). This problem is exacerbated for children and adolescents who place a higher value on present satisfaction while more heavily discounting future consequences. Finally, financial ‘externalities’ exist in the market for SSBs - consumers do not bear the full costs of their consumption decisions. Given the contribution of SSB consumption to obesity as well as the health consequences independent of weight, SSB consumption generates excess health care costs. Medical costs for overweight and obesity alone are estimated to be 9.1% of U.S. health care expenditures ($147 billion), with half of these paid for publicly through Medicare and Medicaid.31
Key issues for an effective policy include which beverages would be taxed, the type of tax (sales vs. excise tax), and the tax rate. We propose a one cent excise tax per ounce for beverages with any caloric sweetener added. An alternative would be to tax beverages that exceed a threshold of grams of added caloric sweetener or kcal per ounce. If this approach were used we would recommend the threshold be set at 1 gram of sugar per ounce (32 kcal per 8 ounces). Another option would be a tax assessed per gram of added sugar, but such an approach would be difficult to administer. There is virtue to taxing beverages with any sugar added because of simplicity and the possibility of promoting consumption of no-calorie beverages, most notably water, but a threshold approach would also promote calorie reductions and encourage manufacturers to reformulate products. A consumer drinking a conventional soft drink (20 ounces) per day and switching to a beverage below this threshold would consume approximately 174 fewer calories each day.
A specific excise tax (levied on units such as volume or weight) per ounce or per gram of added sugar would be preferable to a sales tax or an ad valorem excise tax (levied as a percentage of price) and would provide an incentive to reduce the sugar for each ounce of an SSB. Sales taxes added as a percentage of retail cost have three disadvantages; a) they could simply encourage purchase of lower-priced brands (offering no calorie reduction) or large containers that cost less per ounce; b) they are seen by consumers only after a purchase decision has been made; and c) they would leave untaxed syrups used in fountain drinks with multiple refills. A number of states currently exempt SSBs from sales taxes along with food, presumably because food is a necessity. This anachronism should be eliminated whether or not an excise tax is enacted.
Excise taxes can be levied on producers and wholesalers and would likely be passed through and hence incorporated into the retail price and thus be seen by consumers at the point of purchase. Taxes levied on producers and wholesalers are far easier to collect and enforce than those levied on retailers because many fewer businesses must comply, and the sugar in syrups can be taxed, a major advantage because of the heavy sales of fountain drinks. Experience with tobacco and alcohol taxes suggests that specific excise taxes have a greater impact on consumption than do ad valorem excise taxes and also can generate more stable revenues by virtue of being less dependent on industry pricing strategies.32 In addition, tax laws should be written to adjust specific excise taxes regularly to keep pace with inflation in order to prevent their impact on both prices and revenues from eroding over time.
A tax of one cent per ounce of beverage would increase the cost of a 20-ounce soft drink by 15–20%. Impact on consumption can be estimated with research on price elasticity and by incorporating both direct and indirect effects of price changes. The price elasticity for all soft drinks is in the range of −0.8–1.0.33 Even greater price effects are expected from taxing only SSBs as some consumers will switch to diet beverages. Using a conservative estimate that consumers would substitute 25% of reduced calorie consumption with calories in other forms, a one cent per ounce excise tax would lead to a minimum 10% reduction in calorie consumption from sweetened beverages, or 20 kcal per person per day, sufficient for weight loss and reduction in risk.34 The benefit would be larger in higher-volume consumers, who are more likely to be overweight and who appear more responsive to prices.7 Larger taxes would have greater benefits.
A controversial issue is whether to tax beverages with non-caloric sweeteners. No adverse health effects of non-caloric sweeteners have been consistently demonstrated, but there are concerns that diet beverages may increase calorie consumption by justifying consumption of other caloric foods or by promoting a preference for sweet tastes.35 At present we do not propose taxing beverages with non-caloric sweeteners, but recommend close tracking of future studies to determine whether taxing might later be justified.
The revenue generated from a tax on sugar-sweetened beverages would be considerable and could be used to help support child nutrition/obesity prevention programs, health care for the uninsured, or general revenue needs. A national tax of 1 cent per ounce on SSBs would raise $14.9 billion in the first year alone. Taxes at the state level, as examples, would generate $139 million in Arkansas, $173 million in Oregon, $221 million in Alabama, $928 million in Florida, $937 million in New York, $1.2 billion in Texas, and $1.8 billion in California. A tax calculator is available online that can generate revenue numbers for states and 25 major cities.36
One objection to a SSB tax is that it would be regressive. This argument arose with tobacco taxes but eroded with several realities: the poor face a disproportionate burden of smoking-related illnesses and nearly all smokers begin as teenagers; both groups are sensitive to price changes.7 In addition, some of the tobacco revenue has been used for programs developed specifically for the poor and for youth. The poor are most affected by illnesses related to unhealthy diets, and brand loyalties for beverages tend to be set by the teenage years. In addition, SSBs are not necessary for survival and an alternative (water) is available at little or no cost, hence a tax that shifts intake from SSBs to water would benefit the poor by improving health and lowering beverage expenditures. Designating revenues for programs on child nutrition, obesity prevention, or health care for the uninsured would preferentially help those most in need.
A second objection is that taxing SSBs will not solve the obesity crisis and is a blunt instrument that affects even those who consume low amounts. Seat belt legislation and tobacco taxation do not eliminate traffic accidents and heart disease but are sound policies. Similarly, obesity is unlikely to yield to any single policy intervention, so it is important to pursue multiple opportunities to create incremental gains. Reducing caloric intake by 1–2 per cent per year will make a marked impact on health of all age groups, and the financial impact on those who consume small amounts would be minimal.
Opposition to a tax by the beverage industry is to be expected given the possible impact on sales, has been seen in jurisdictions which have considered such taxes, and can be predicted from behavior of the tobacco industry under similar circumstances.37 PepsiCo threatened to move its corporate headquarters from New York when the state considered implementing an 18% sales tax on SSBs.38 The tobacco industry fought policy changes by creating front groups with names suggesting community involvement. The beverage industry has created Americans Against FoodTaxes.39 These reactions suggest that the beverage industry believes a tax would have a substantial impact on consumption.
Public support for food and beverage taxes to address obesity has increased steadily. Polls have asked about taxes in different ways and hence are not directly comparable from year to year, but overall trends are clear. Support for food taxes rose from 33% in 2001, to 41% in 2003, and then to 54% in 2004.40 More recent polling shows higher support for beverage taxes. A 2008 poll of New York State residents found that 52% support a soda tax, rising to 72% when the revenue would be dedicated to programs to prevent obesity in children and adults. How the issue is framed is essential, with highest support when the tax is introduced in the context of promoting health and the revenues earmarked for child nutrition or obesity prevention programs.
The federal government, a number of states and cities, and some countries (e.g., Mexico8) are considering taxes on sugar-sweetened beverages. The reasons to proceed are compelling. The science base linking sugar-sweetened beverage consumption with risk for chronic diseases is clear. Escalating health care costs and the rising burden of diseases related to poor diet create an urgent need for solutions, thus justifying government’s right to recoup costs.
As with any public health intervention, the precise impact of a tax cannot be known until it is implemented and studied, but research to date suggests strong positive effects on reducing consumption.5, 33, 34 In addition, the tax has the potential to generate substantial revenue to prevent obesity and address other external costs resulting from SSB consumption, as well as fund other health-related programs. Much as tobacco taxes are routine at both state and federal levels because of revenue generation and the public health benefit on smoking rates, we believe taxes on beverages that help drive the obesity epidemic should and will become routine.
Work on this paper was supported in part by grants from the Rudd Foundation (to KDB), the National Institutes of Health (to BMP; NIH R01-CA121152), and the Robert Wood Johnson Foundation (to FJC).
All authors report no conflict of interest relevant to this article.
Dr. Kelly D. Brownell, Rudd Center for Food Policy and Obesity at Yale University.
Dr. Thomas Farley, New York City.
Dr. Walter C. Willett, Department of nutrition at the Harvard School of Public Health.
Dr. Barry M. Popkin, Department of nutrition and director of the UNC Interdisciplinary Obesity Center at the University of North Carolina, Chapel Hill.
Dr. Frank J. Chaloupka, University of Illinois at Chicago and director of the UIC Health Policy.
Dr. Joseph W. Thompson, University of Arkansas for Medical Sciences and serves as the Surgeon General for the State of Arkansas.
Dr. David S. Ludwig, Optimal Weight for Life program at Children’s Hospital Boston and associate professor of pediatrics at the Harvard Medical School.