Should We Fight Today’s War on Obesity Like the Last War on Tobacco?
Friday, August 7, 2009
Moderation in one’s personal habits should be matched by moderation (or at least more humility) in public policy.
The health consequences of obesity in the U.S. population remain a serious and growing problem. However, government officials, health experts, and the public still have trouble sizing it up accurately, let alone getting their hands around it (“Take my waist line, please!”). The latest set of proposals on this front arrived late last month in an Urban Institute report, “Reducing Obesity: Policy Strategies from the Tobacco Wars.”
Co-authors Carolyn Engelhard, Arthur Garson Jr., and Stan Dorn recommend three primary policy interventions to modify the environment in which people make food choices, so that healthy choices become easy and decisions to eat fattening food become more difficult. They suggest that policy tools that helped bring down tobacco use since the mid-1960s could be modified and applied to cut the fat and fight obesity today.
In particular, Engelhard, Garson, and Dorn urge new taxes on fattening food of little nutritional value, putting simpler nutritional labels for consumers on the front of packaged foods and on restaurant chain menus, and banning advertising and limiting the marketing of fattening food. (The last two recommendations could be translated as “government information good; private market information bad.”)
Whether a tax on fast food and junk food is a good idea depends mainly on the extent to which individuals already pay fully for the consequences of their decisions about diet and exercise.
But the first round of discussion of the new report overlooked recent research challenging several of its premises and suggesting unintended consequences ahead. For example, the authors claim that the health costs of obese individuals impose a serious market externality on the rest of us, including raising private premiums for non-obese workers nearly $26 billion higher each year. They base this estimate on formulaic indexing and proportional allocation of per-capita health spending costs attributed to obesity, originally based on 1998 data as previously estimated by researchers Jay Bhattacharya and Neeraj Sood in a 2005 National Bureau of Economic Research (NBER) working paper. Bhattacharya and Sood’s paper also had concluded that, in the absence of a legally authorized risk rating for observable risk factors such as obesity, the failure of the obese to pay for their higher medical care expenditures through higher health insurance premiums in employer plans that pool health risks could reduce incentives for those individuals to maintain a normal weight.
However, Bhattacharya continued and extended this research, with his Stanford University colleague Kate Bundorf. In the May 2009 issue of the Journal of Health Economics, they find that the incremental healthcare costs associated with obesity actually are passed on to obese workers with employer-sponsored insurance (ESI), in the form of lower cash wages. Notably, this wage offset for the health costs of obesity does not occur for workers without ESI coverage. It also appears to be less significant for obese male workers in employer plans, because their medical expenditures are not generally higher than those of their normal-weight counterparts in a firm (perhaps men have more front-end “carrying” capacity?). Bhattacharya and Bundorf estimate that obese men earn $1.21 an hour less than non-obese men, while obese women earn $1.66 less than non-obese women. However, for women, the obesity penalty is even higher in firms where their employers provides health insurance—they earn $2.64 an hour less. Bhattacharya and Bundorf point out that whether a tax on fast food and junk food is a good idea depends mainly on the extent to which individuals already pay fully for the consequences of their decisions about diet and exercise. But in the case of obesity, wage offsets do adjust for such weighty variations in the individual characteristics of workers that would affect their expected medical expenditures and make them high-cost to insure. It turns out that there is less real pooling for those workers, who are more likely to sink to the bottom than float. And new “twinkie” taxes would not only distort the decisions and tradeoffs already made in labor markets; it might just subject high-calorie (and likely lower-income) consumers to another form of super-sized “double taxation”—paying once through lower wages and again through added excise taxes; along with poorer health later in life!
Moreover, other NBER work published just last month by Bhattacharya, Bundorf, Sood, and Noemi Pace indicates that insurance coverage does encourage the wrong behaviors related to body weight, by insulating people from the costs of obesity-related medical care expenditures. However, those effects are larger in public insurance programs, where premiums are not risk-adjusted, and smaller in private insurance markets where the obese are more likely to pay incremental medical care costs in the form of lower wages. Note that the current thrust of health "reform" proposals before Congress would be to increase such public coverage at the expense of private coverage. It will be hard to provide incentives for healthy behaviors (despite some political lip service directed to that goal) if most of the new money for expanded coverage points in the opposite direction.
Passing a new law alone often is not equivalent to achieving its stated objectives. Particularly when the behavioral problem for most obese persons is not that they do not know they are overweight or that what they are eating might contribute to that condition.
To recap, the insurance-related incentives and resulting costs regarding obesity are complex and operate at several levels. Simply having any kind of health insurance diminishes an obese individual’s consciousness of the full economic costs of their excessive weight. However, higher levels of coinsurance dampen this effect. Moreover, the source of insurance matters. Public program coverage tends to hide the costs of obesity more completely (passing them along to taxpayers, with both minimal cost sharing and community-rated premiums for beneficiaries). Employer sponsors of private group coverage shift most of those costs back on to obese workers (particularly obese women) through wage offsets that lower the latter’s earnings. Individual insurance market coverage provides even stronger incentives to avoid or limit obesity (more cost sharing and risk-rated premiums). Although Bhattacharya and Bundorf do not speak to this point, it is reasonable to conclude further that paying higher health premiums for being obese would provide a stronger and more transparent incentive to reduce one’s weight than just experiencing the less direct “penalty” of lower wages over time.
The authors of the Urban Institute report also rather blithely assume that the proceeds of any new taxes on fattening foods could be effectively allocated to finance anti-obesity activities, subsidize better food choices for low-income households, and fund “national health reform.” They forget how the higher cigarette taxes of the late 1990s soon were converted largely into rather leaky “rainy day” funds in state budgets that soon served more compelling political claims than those of anti-smoking advocates. This year, the strongest political advocates of new taxes on sugary soft drinks and fattening foods are those desperately seeking to find more revenue to finance expansions of insurance coverage toward near-universal levels. If the bait of raising revenue for purported anti-obesity objectives is taken, the switch to spending it on those larger strategic goals will not be far behind.
Engelhard, Garson, and Dorn place renewed faith in new and improved nutrition labels to reduce consumption of fattening foods. Those newer models of simplified labels will have to work better than the older ones, launched under the Nutrition Labeling and Education Act (NLEA) of 1991. NBER researchers concluded in a 2006 review that the benefits of the NLEA labels were limited to certain demographic groups (mostly non-Hispanic white females). Most groups in the United States did not seem to benefit (at least in terms of reduced body weight) from these regulations. This serves as one more reminder that passing a new law alone often is not equivalent to achieving its stated objectives. Particularly when the behavioral problem for most obese persons is not that they do not know they are overweight or that what they are eating might contribute to that condition.
The larger lesson is the importance of recognizing tradeoffs in costs and benefits across the entire range of human behavior before launching one-dimensional policy crusades whose reach exceeds their grasp.
One also could question the Urban Institute report’s over-reliance on measures of body mass index (BMI) as the appropriate marker for obesity levels. BMI fails to distinguish fat from fat-free mass such as muscle and bone, as opposed to better measures such as percent body fat, waist circumference, or the ratio of waist size to hip size. (See here and here.)
Finally, it is important to recall the law of unintended consequences that often operates when public policy tries to narrowly reshape any one particular social behavior that is interconnected with other ones, and that is often driven by broader social trends. Research earlier this decade by Michael Grossman, Shin-Yi Chou, and others shows that eating out at fast-food restaurants and full-service restaurants may be the most important factor in explaining the rise in obesity—followed by substantial increases in the real price of cigarettes. And the main causes behind those trends were the increased rates of labor force participation by women and the crackdown on smoking via tax increases. The first factor involves the increasing scarcity and increasing value of household time outside the labor market. Time devoted to at-home meal preparation decreased, and the microwave-ready or drive-thru alternatives gained a large share of the market (and circumference-adjusted waistlines). The second factor involves higher cigarette taxes that not only caused more smokers to quit, but to begin eating more as a result. If only Mrs. Cleaver had stayed at home, bought more cookbooks, and read them while puffing away! At least she would not have had to buy a new wardrobe.
The larger lesson is the importance of recognizing tradeoffs in costs and benefits across the entire range of human behavior before launching one-dimensional policy crusades whose reach exceeds their grasp. Moderation in one’s personal habits should be matched by moderation (or at least more humility) in public policy. Otherwise, recent research on the socially contagious “network effects” of obesity might also lead to new federal restrictions on one’s friends and family, who do not measure up to BMI limits. Let’s not refight the last war with the crude weapons of mass consumption used by the last generation of politicians and public health generals.
FURTHER READING: Miller wrote “What You Don’t Know Can Hurt You,” on how there is a strong association between educational attainment and health—one more reason to empower Americans, not Washington, with greater ownership of their healthcare. His other articles include “What DO We Know About the Uninsured?” and “Obama Healthcare 2.0,” on how the president’s opening offer of healthcare at a teaser rate fails to deliver what we actually need, value, and can afford.
Image by Darren Wamboldt/Bergman Group.