2009-0721 364.37

doi: 10.1377/hlthaff.2009.0721HEALTH AFFAIRS 29,NO. 3 (2010): 364–3712010 Project HOPE— The People-to-People HealthFoundation, Inc.
ABSTRACT In the past few decades, obesity rates among American children .edu) is an associate professorin the Department of Policy have skyrocketed. Although many factors have played a part in this Analysis and Management atCornell University in Ithaca, unhealthy increase, this paper focuses on how economic policies may be contributing to our children’s growing girth and how these policiesmight be altered to reverse this trend. It examines the economic causesand consequences of obesity, the rationales for government intervention,the cost-effectiveness of various policies, and the need for more researchfunding.
Inthetwenty-fiveyearsthatelapsedbe- discussingtheimplicationsforhealthpolicy.
Economic explanations for the recent rise in obesity tend to focus on changes that give people incentives to consume more or to burn fewer calories. These include the following.
4 percent to 15.3 percent among six-to-eleven- FOOD PRICES The real price of food (the price of year-olds; and from 6.1 percent to 15.5 percent food adjusted for inflation across all goods and services) has declined greatly in recent decades.
For example, between 1990 and 2007 the realprice of a two-liter bottle of Coca-Cola fell34.89 percent, and that of a McDonald’s quar-ter-pounder with cheese fell 5.44 percent.3 Some research indicates that reductions in the price of food account for 41–43 percent of the rise in Childhood obesity is a complex, multidiscipli- young adults’ body mass index (BMI) between nary issue, so the search for solutions requires information sharing across many fields. To pro- BMI is most sensitive to the price of fast food in mote such exchange, this paper describes re- particular for youth from families of low socio- search findings from the field of economics.
economic status.5 The real price of fruit and veg- Relative to other disciplines, economics has only etables rose 17 percent between 1997 and 2003,6 recently been applied to the study of obesity. As a an increase that some studies have linked to result, there are many useful applications yet to Economics offers several useful tools for the AGRICULTURAL POLICIES The decline in the price study of childhood obesity: insights into the eco- of food could be due to agriculture policies. How- nomic causes and consequences of obesity, ever, several studies have concluded that such clearly defined rationales for government inter- policies have little impact on consumer prices of vention in markets, and the use of cost-effective- energy-dense foods.9,10 About half of farm subsi- ness analysis for comparing policies to prevent dies are based on historic, not current, produc- or reduce obesity. This paper summarizes recent tion, which limits incentives for farmers to in- research in each of those areas. It concludes by crease output.9 Moreover, U.S. farm policy raises the prices of sugar (by restricting imports) and milk (through regional dairy compacts).
In addition, the impact of farm policy on many that the poor are more likely to be obese, the retail food prices is modest because the cost of extensive literature on socioeconomic status farm products (ingredients) represents a small and obesity includes widely varying findings.18 proportion of the food’s retail price. For exam- For example, one recent study finds that the cor- ple, the cost of corn—the key ingredient in high- relation between childhood obesity and the fructose corn syrup—contributes just 1.6 percent income of the household varies considerably to the price of soft drinks.11 It is estimated that by race, sex, age, and year. The study also found agriculture subsidies, by decreasing the prices of a negative correlation between childhood obe- food, raise the BMI of youth by 0.08 percent, or a Subsidizing the production of corn and re- may contribute to childhood obesity. It is esti- stricting imports of sugar have given food man- mated that the increase in a mother’s average ufacturers, particularly the producers of soft weekly work hours explains 11.8–34.6 percent drinks, an incentive to substitute high-fructose of the rise in childhood obesity in high-socio- corn syrup for sugar.11 Some researchers have economic-status families between 1975 and argued that this shift contributes to increased 1994.20 A subsequent study found an association food intake and obesity because fructose, unlike between maternal employment and childhood the glucose in sugar, does not stimulate insulin obesity only for children whose mothers were secretion or leptin production, both of which employed full time when the child was seven regulate food intake.13 However, the link be- years old. This association was not found for tween high-fructose corn syrup and obesity children whose mothers worked full time when the child was either preschool age or age eleven, Another way in which agriculture policy con- or for children whose mothers worked part time tributes to increased food consumption and thus obesity is that the U.S. government requires pro- ducers of agricultural commodities that enjoy television than children whose mothers do not price supports to contribute a specific amount work outside the home.22 In addition, working of money for each unit they sell into a fund that is mothers spend less time cooking and eating with used for commodity-specific advertising and re- their children than mothers who do not work search. Among other things, these funds are outside the home, which increases the probabil- used to support the advertising of fast-food ity that the family will consume prepared foods.
menu items, which in turn raises the risk of These decreases in time spent cooking and eat- ing with children are only slightly offset by INCOME Income is another potential contribu- increases in time spent at such activities by hus- tor to obesity. Higher income could prevent weight gain by enabling consumers to substitute Child care arrangements for the offspring of healthier, more costly food for cheaper energy- working mothers have also been correlated with dense food or by increasing demand for good an increased risk of childhood obesity. Children health or an attractive appearance.15 On the in center-based care are more likely to be obese other hand, additional income could promote than children cared for by nonparental relatives weight gain by allowing people to consume more or by a nanny.24 This may be attributable to differ- calories and spend more time in sedentary pur- ences across the environments in the quality of suits. These conflicting effects may explain why research so far has found little or no evidence ing alcoholic beverages, but research has notborne that out. Studies find only about a 10 per- cent correlation among these behaviors.28 obesity include labor-market costs to the Economic Consequences OfChildhood Obesity The economic consequences of childhoodobesity are typically categorized as direct, such as medical costs, and indirect, such as job absen-teeism. The direct costs of childhood obesity in- clude annual prescription drug, emergencyroom, and outpatient costs of $14.1 billion,29 plusinpatient costs of $237.6 million.30 An even larger cost is incurred when obese children become obese adults. About a thirdof obese preschool children, and about half The government could tax obesity itself (which of obese school-age children, become obese would presumably be politically unpopular), or adults.31 The estimated annual cost of treating tax the behavior that contributes to obesity and obesity-related illness in adults is $147 billion.32 subsidize behavior that contributes to healthy The medical costs of obesity are so substantial that the rise in obesity explains 27 percent of the Another government intervention that might rise in health care spending between 1987 and be useful is to provide consumers with more 2001.33 A limitation of this research is that information about the food they eat. Recently, obesity may be correlated with unobservable fac- cities such as New York have required fast-food tors that directly affect medical costs; for exam- chain restaurants to list the calorie content of ple, a person may have sustained a disabling their food on menus and menu boards.38 It is not injury that leads to obesity as well as raises medi- yet clear whether such menu labeling affects food choices and, ultimately, obesity.
The indirect costs of obesity include labor- More is known about the effects of the Nutri- market costs to the obese individual and poten- tion Facts panel, which has been required on tially the employer. Obesity is associated with packaged foods since 1994. Since the Nutrition delayed skill acquisition in children as young Facts panel was introduced, Americans have in- as two to three years old.34 Even when education creased consumption of iron and fiber, although and cognitive ability are controlled for, adult they have not lowered consumption of total fat, obesity is associated with lower wages, especially for white females.35 In addition, obesity-related Obesity prevalence among white females who job absenteeism totals $4.3 billion annually.36 use the labels when shopping is 2.67 percentage Also, obesity is associated with lower productiv- points lower than it would have been in the ab- ity while at work (presenteeism), which totals sence of the labels (no significant effect of the labels on weight was found for other demo-graphic groups).40 Another economic rationale for government intervention in markets is to protect consumers who are acting irrationally. This criterion should One economic rationale for government inter- be used cautiously, because irrationality is in the vention is to internalize external costs, which means to ensure that all of the costs associated However, children arguably are not able to with obesity are borne by those whose decisions take into account the future consequences of create the costs. In 2008 obesity-related illness their actions, and children as old as age eleven cost Medicare $19.7 billion and Medicaid, $8 bil- may fail to appreciate the difference between lion.32 These costs were borne by the general television advertisements and television pro- population, whose tax dollars fund these federal gramming.41 As a result, some European coun- and state health insurance programs. In addi- tries have banned food companies from adver- tion, private health insurance plans paid $49 bil- tising to children.41 It is estimated that a ban on lion to treat obesity-related illness in 2008.32 television fast-food advertising to children in the Some of this cost also was borne by the nonobese United States would reduce the prevalence of in the form of higher group health insurance overweight by 18 percent among children ages 3–11 and by 14 percent among youths ages 12– 18.42 However, it is difficult to accurately mea- cure” is true in this case. Treating adult obesity sure the effect of food ads on weight because with the drug orlistat costs $8,327 per QALY,46 exposure to the ads is correlated with sedentary the drug sibutramine costs $9,299 per QALY.47 lifestyle and other factors that contribute to Gastric bypass surgery costs $5,000–$16,100 per QALY for women and $10,000–$35,600 perQALY for men.48 Wheeling Walks, a behaviormodification program targeted to sedentary adults, costs $14,286 per QALY; other behavior To get the biggest “bang for the buck” in prevent- modification programs for adults cost much ing childhood obesity, decision makers can use cost-effectiveness analysis, which compares in- Of course, not every prevention strategy is terventions in terms of their cost per quality- more cost-effective than every treatment. For ex- adjusted life-year (QALY) saved.43 Exhibit 1 lists ample, the Moving School Bus program, which some of the available data on the cost-effective- encourages children to walk rather than ride to school, is neither effective nor cost-effective.50 Based on the published literature, the most Limitations of these cost-effectiveness analy- cost-effective way to prevent obesity in youth ses include many uncertainties, such as how is the Coordinated Approach to Child Health effectively participants maintain weight loss (CATCH), a comprehensive intervention to pro- mote healthy eating and physical activity inelementary schools, which costs $900 per QALYsaved.44 The next most cost-effective program is Planet Health, a comprehensive intervention to Public opinion polls indicate that most Ameri- promote healthy eating and physical activity in cans consider childhood obesity to be a major middle schools, which costs $4,305 per QALY public health problem.51 In a 2008 nationwide saved for girls but is not effective for boys.45 survey, obesity was ranked as the number-one The adage that “prevention is cheaper than Cost Per Quality-Adjusted Life-Year (QALY) Saved Of Various Interventions To Prevent Or Reduce Obesity walking, frequent calls to promptparticipants to walk However, polls may be misleading indicators of taxpayers’ “willingness to pay” higher taxesfor obesity prevention programs. So-called con- tingent valuation survey methods, which areused to estimate the value of goods not traded in markets, indicate that although 83 percent ofNew York State residents agree that childhood obesity is a major problem, more than a third refuse to pay even $10 a year in higher taxes to cutchildhood obesity in half.53 Collectively, how- ever, society is willing to pay substantialamounts to reduce childhood obesity. The meanwillingness to pay for a 50 percent reduction inchildhood obesity in New York State was $46.41,which implies a total willingness to pay by NewYork State residents of $690.6 million.53 The level of public support for anti-obesity these differences is statistically significant policies is greatly influenced by how the issue (p < 0:01). Thus, how the issue of costs is framed of costs is framed. The 2009 Empire State Poll of can cut support by more than half, from 92.1 per- New York State residents asked respondents cent to 40.5 percent, and can increase opposition whether they agree with improving the nutrition twelvefold, from 4.4 percent to 53.2 percent.
of food in school cafeterias (Exhibit 2) and in- Similar patterns are shown in Exhibit 3 for creasing the quantity and quality of physical ed- questions about increasing the quantity and ucation in schools (Exhibit 3). For each of these quality of physical education in schools. Clearly, policies, respondents were randomly asked one there is a need for caution when using poll of three versions of a question that varied in how results to gauge public support for specific explicitly costs of the policy were discussed.
policies—support may be highly sensitive to Exhibit 2 shows that if no mention is made of how questions are worded, especially with re- costs, 92.1 percent agree and only 4.4 percent disagree with improving the nutrition of food inschool cafeterias. If the question is modified toend “even if it requires raising taxes,” agreement falls to 69.5 percent and disagreement quintu- Inevitably, government efforts to prevent and ples, to 22.6 percent. If the question begins with reduce childhood obesity will have limited bud- a discussion of costs by saying “The government gets. To maximize the benefit to society, it is should raise taxes in order to improve the nutri- necessary to spend that limited budget on the tion of food in school cafeterias,” only 40.5 per- anti-obesity interventions that are most cost- cent agree and 53.2 percent disagree. Each of How The Framing Of Costs Influences Public Support For Improving The Nutrition Of School Food The government should improve the nutrition of food in school The government should improve the nutrition of food in school cafeterias, even if it requires raising taxes The government should raise taxes in order to improve the nutrition of SOURCE Author’s calculations, using data from the 2009 Empire State Poll. NOTES The 2009 Empire State Poll (ESP) of 800 respondentswas conducted 30 January–16 March 2009. Respondents were selected using the random-digit-dial method, and the interview wasconducted using a computer-assisted telephone interviewing software system. The random sampling frame allows poll results tobe generalized to the entire state of New York. Percentages do not add to 100 because respondents also had the option ofanswering that they were “neutral,” answering “don’t know,” or refusing to answer. In the second and third rows, the portion of thequestion that is underlined indicates the part of the question that differs from the wording in the first row. It does not indicatethat the interviewer emphasized those words when asking the question. The difference in values across rows within a column isstatistically significant (p < 0:01) in each case.
How The Framing Of Costs Influences Public Support for Improving Physical Education In Schools The government should increase the quality and quantity of physical The government should increase the quality and quantity of physical education in schools, even if it requires raising taxes The government should raise taxes in order to increase the quality and quantity of physical education in schools SOURCE Author’s calculations, using data from the 2009 Empire State Poll. NOTES: See Exhibit 2 SET PRIORITIES For example, the school-based provide a rationale for taxes and subsidies to program CATCH is highly cost-effective.44 The promote healthy eating. Experiments have es- wide variations in cost-effectiveness shown in tablished that altering food prices can change Exhibit 1 underscore the importance of prioritiz- youth behavior. Cutting the prices of fruit, salad, ing interventions with the lowest cost per QALY, and carrots in high school cafeterias in half re- to avoid wasting resources. Unfortunately, cost- sulted in a quadrupling of sales of fruit and a effectiveness studies have not yet been con- doubling of sales of baby carrots, although it did ducted for many promising interventions. Re- not significantly affect sales of salads.58 Other search is, in many cases, at the earlier stage of experiments with vending machines in high determining whether the intervention is effec- schools found that the quantity and type of tive at all. The National Institutes of Health snacks purchased were significantly affected by (NIH) should continue to invest in cost-effective- price.59 A limitation of these studies is that they ness studies of anti-obesity interventions, to bet- do not track consumption in all places and at all ter inform funding decisions and ensure that the times. Also, it remains unclear whether total limited funds allocated to obesity prevention and caloric intake rises or falls when low-calorie treatment yield the maximum improvement in A challenge for food taxes is defining the scope of what should be taxed and what should be mandate that private health insurance plans subsidized. Should apple juice be taxed because cover cost-effective methods of preventing and it is energy-dense, or should it be subsidized treating childhood obesity.55 These policies may because it is 100 percent fruit and sometimes not be covered in the absence of a mandate be- contains added vitamins? Prominent public cause high turnover of enrollees implies that it is health advocates have recently called for taxes profit-maximizing for health insurance compa- on full-calorie soda.60 The mean published esti- nies to refuse to reimburse for prevention and mate is that a 1 percent increase in soda price treatment that do not generate short-term sav- would reduce the quantity of soda demanded by ings. As a result, many health insurance plans do 0.79 percent.61 The modest soft drink taxes that not reimburse doctors for preventing and treat- currently exist appear to have negligible effect on body weight.62 A challenge for estimating the impact on obesity of a substantial soft drink health insurance programs should cover cost- tax is that no such taxes currently exist.
effective methods of prevention and treatment.
A risk of taxing a narrow set of energy-dense West Virginia and Pennsylvania have been in- foods (such as soda) is that consumers may “sub- novators in addressing obesity by expanding stitute away” from the taxed items toward Medicaid coverage—for example, for nutrition energy-dense items that are not taxed. Another education.57 A handful of states are incorporat- challenge is that food taxes can be regressive, ing incentives for healthy behavior into their falling more heavily on the poor. Several states Medicaid and Children’s Health Insurance Pro- that tax food, including Hawaii, Idaho, Kansas, Oklahoma, South Dakota, and Wyoming, both praised for encouraging healthy behavior and tax food and provide an income tax credit or criticized for placing the health of low-income rebate to low-income households that compen- children at risk by linking the provision of serv- sates them for the loss of their purchasing USE INCENTIVES The external costs of obesity search and policy can help identify contributors Obesity is a major public health problem facing to obesity, calculate the consequences of obesity, the United States, with significant consequences and allocate scarce resources to the interven- for health care costs and productivity. Incorpo- tions that offer the greatest benefit per dollar rating the economic perspective into obesity re- Johnson Foundation’s Scholars in Health 3 Christian T, Rashad I. Trends in U.S.
tion and costs in childhood. Obesity.
hospital care and costs, 1999–2005.
payer- and service-specific estimates.
34 Cawley J, Spiess CK. Obesity and skill policy: fact or rhetoric? Food Policy.
26 Cutler DM, Glaeser EL, Shapiro JM.
obesity. Am J Med. 2002;113:491–8.
choices. J Nutr. 2003;133:841S–3S.
(NY): Oxford University Press; 1996.
implications for taxes and subsidies.
tients with type 2 diabetes mellitus.
63 Federation of Tax Administrators.
to do with it? Health Aff (Millwood).

Source: http://cfapps.sph.uth.tmc.edu/uploadedFiles/Redesign_Website/Research/Research_Centers/Michael_and_Susan_Dell_Center_for_Advancement_of_Healthy_Living/Economics%20of%20Childhood%20Obesity.pdf