Thursday, April 03, 2014

Obesity and Taxes

Cigarette taxes have proved to be an effective strategy to reduce smoking, so one might think (as many experts do) that soda taxes would be an effective strategy to reduce obesity. Consumption of soft drinks seems to be an important risk factor for obesity, and people are sensitive to the price of their colas.

Moreover, soda taxes reflect the lessons of behavioral economics. People often over-indulge in activities that provide short-term gratification but impose significant harm in the long-term. Imposing a tax on unhealthy drinks supplies an immediate disincentive to the consumption of those drinks and can overcome the difficulty people have in postponing gratification. 

However, two new articles in Health Economics question the effectiveness of soda taxes. One study based on UK data, the other on US data, come to the same conclusion--we should not expect much of an impact from taxes on sugary soft drinks. It seems that raising taxes on some beverages simply results in consumers switching to other beverages and replacing the forgone calories with other calories. And to make things worse, beverage taxes often are regressive.

The news is disappointing and adds to a growing list of disappointing policies for weight loss. Under the Affordable Care Act, for example, restaurants must disclose calorie information to customers. With better information, diners would know which salads really are healthy and which others are not. But researchers have not found mandates for calorie disclosure by restaurants (as in New York and Seattle) to be effective.

Legislative fixes for obesity are tempting and probably necessary. But lawmakers need to take better account of medical understanding before they act.

[cross-posted HealthLawProfs and orentlicher.tumblr.com]

Posted by David Orentlicher on April 3, 2014 at 09:00 AM | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Obesity and Taxes:



The problem with your argument, from my perspective, is that these policies have not been billed as generic consumer rights initiatives to my knowledge, but as public health measures (hence the menu provision's inclusion in the Affordable Care Act, to be implemented by FDA regulation). With evidence indicating that they fail as health measures, we now get an altogether different rationale.

A "rights" rationale seems strained. No one has a right to eat out. If consumers don't like what is on offer at a given establishment, including a dearth of nutritional information, they are free to vote with their dollars. So the question for me seems less a matter of rights than what we get for a policy that everyone acknowledges has significant costs.

If all we get is knowledge that the subset of the population who needs it disregards, it is not worthwhile by my lights (and I say that as someone who actually likes the additional information). I think you are right about "importance of the information," but it is only important because of its public health implications, not because of a consumer's right to know in the abstract.

Posted by: The Curmudgeonly Ex-Clerk | Apr 5, 2014 12:42:35 AM

@CEC - I think the larger issue is that one need not necessarily rely on an assumption/justification that calorie information will change the behavior of the obese. Rather, there are other benefits to be served, like a consumer's right to know. Michelle is correct that it can be costly to present the information on menus versus websites, but given the importance of the information, I would be inclined to make it as accessible as possible.

Posted by: Susannah Pollvogt | Apr 4, 2014 4:22:08 PM


Like you, I appreciate and use nutritional data to make decisions (sometimes). But I'm skeptical that the kind of folks who hang out in the PrawfsBlawg comments section (likely well-educated members of a certain social class with analytical skills honed by legal or other specialized training) are representative of the demographics most in need of better nutrition. I suspect the people most likely to make use of additional nutritional data, which already is plentiful, are the people who least need it; those who have some interest in eating healthy already take steps to do so.

I think many assume that, if only the obese were better informed, they would be non-obese or at least less so. But whatever specific data the obese might be lacking in connection with menus, surely they understand that their diets and lack of physical activity have something to do with their belt sizes. Yet, they persist in their dietary habits. The assumption underlying mandatory menu disclosures seems to be that large calorie counts will hold a persuasive value that large belt sizes (and the omnipresent reality of being obese) lacks. Isn't that dubious on its face?

Posted by: The Curmudgeonly Ex-Clerk | Apr 4, 2014 3:09:03 PM

What level of proof should be required before a policy is implemented is a complex question, but I can't agree that a policy about which we can only say that it may "potentially, theoretically facilitate good behavior" suffices when that policy will cost millions of dollars -- and when there's some evidence to suggest that the policy will make the public health problem it's intended to redress worse!

Of course a different policy rationale -- such as the consumer right-to-know rationale that Adam and Susannah raise -- may well justify the cost of the policy. But it's worth being clear which policy rationale we're invoking. For one thing, as I said before, there's a real danger of crying wolf if we implement behavioral interventions on public health (or similar) rationales that turn out to have no effect on a problem or even exacerbate it. Play fast and loose with what counts as "evidence-based policy" or "science-based interventions" and you lose public support for actual evidence-based policy, which is too important to risk.

Second, meeting different rationales may require different implementations of a policy. For instance, if the policy rationale of providing caloric info is that consumers have the right to know what they're putting in their bodies, in the age of smart phones, restaurants can arguably meet this goal by providing nutritional info on their websites and/or in point-of-sale handouts (which many already do), which I'd think would be hugely and modestly cheaper, respectively, than altering all of their menus and menu boards. Only the public health rationale requires calorie counts to be displayed right next to the item -- the hypothesis being that the calorie count would helpfully frame the choice to buy a Starbucks muffin the size of your head with 8 millllllion calories in it. For those like yourself, already "inclined to be proactive," something much more cost effective might well suffice.

And no, the studies I have in mind of calorie menu laws didn't measure the effects of those laws on third parties, because that wasn't the hypothesis. (In fact, because these laws were implemented without testing their efficacy, researchers were limited to observational studies rather than RCTs.) There has been some work on that front (see ‎Christakis on the spread of obesity in social networks) and I'd certainly be interested in hearing about laws and policies designed to leverage that effect. But again, rather than anecdote and intuition, I'd like to see data, and I'd like to see any resulting policy tailored at least somewhat to both the data and the policy rationale.

Posted by: Michelle Meyer | Apr 4, 2014 12:10:56 PM

Agreeing with Adam. Doesn't it make sense to implement public policy on the basis that it can potentially, theoretically facilitate good behavior, rather than requiring that it is "proven" to produce good behavior? I definitely want, need, and make use of nutritional information, and maybe I persuade others to be concerned as well. I don't know if isolated studies can really account for how people who are inclined to be proactive influence the choices of others who are not.

Posted by: Susannah Pollvogt | Apr 4, 2014 11:40:35 AM

Curmudgeonly Clerk:

I thought the calorie count laws (which I see all the time in my city) made sense because, well, why not tell consumers what they are eating? I have been frequently astonished at the high calorie counts. No one thinks a muffin is health food, but 800 calories? No one actually thinks a salad can be bad, and it's very easy to misjudge what "a lot" of dressing looks like. That's a quiz I think a lot of people would get wrong. So, I can't agree with you that the idea this might shape behavior was crazy. (It certainly killed my appetite for Starbucks treats).

However, as Michelle notes, the evidence is equivocal to negative on the calorie count-reducing benefits of these things. I still think I have the right to know, but we can't be very confident everyone will forsake their apple fritters, as I have regretfully felt compelled to do.


Posted by: Adam | Apr 4, 2014 8:05:36 AM

Why would anyone expect calorie disclosure in restaurant menus to be effective? Calories, fat, and other nutritional information are listed on the packaging of most items in the supermarket. Obesity in the face of such disclosure suggests that significant portions of the public either disregard this data in favor of other considerations or are ineducable about such matters.

Even without menu-related disclosures, the longstanding availability of such information in supermarkets provided at least some frame of reference for dining out. For example, anyone who has read the backs of salad dressing bottles understands that a salad accompanied by a liberal pour of most dressings is going to have fat and calories comparable to items on the menu that are recognizably unhealthy.

Posted by: The Curmudgeonly Ex-Clerk | Apr 3, 2014 1:56:37 PM

Matt, I agree that *researchers* shouldn't necessarily give up on these behavioral interventions so quickly. They should try tweaking the choice environment, and so on. Even state and local governments might experiment with these policies (preferably in the true sense of "experiment," i.e., an RCT). But I think it's fair to ask whether it made sense for the ACA to require menu calorie counts across the entire country, effectively ending the possibility of gathering further evidence from which causation can plausibly be inferred. The ACA included the menu regulation in the absence of evidence that it works--and indeed, in the face of some evidence that it not only did not reduce caloric intake but actually increased it.

As I wrote in the LA Times a while back (http://articles.latimes.com/2013/sep/29/opinion/la-oe-chabris-nudge-20130929:

"Implementing untested nudges has real costs. According to the federal government, the Obamacare calorie rule imposes a new 14.5-million-hour paperwork burden, and first-year compliance costs for businesses could total $537 million. If the benefits of a government intervention are not expected to outweigh its costs, then doing nothing will often be the better policy choice."

In addition, implementing regulations that don't work undermines confidence in regulators. Behavioral policies in particular have been attacked from both the right (nannying! mind control!) and the left (neoliberalism!). Those who support a practical role for behavioral economics ought to be more cautious about implementing untested and/or unproven measures, especially on a national scale -- if not for the sake of those regulated, then for the sake of the future of behavioral policy.

Posted by: Michelle Meyer | Apr 3, 2014 11:49:13 AM

But researchers have not found mandates for calorie disclosure by restaurants (as in New York and Seattle) to be effective.

I wonder if we don't often take too short of a perspective on these things, expecting results faster than is reasonable and getting disappointed when we don't see the desired results right away. Wouldn't the real surprise be if people changed their behavior within a fairly short time- by which I mean a few years, even- because of things like this? Of course, it would be nice to be able to expect quick change, and for change to be easily attributable to small actions, but that seems to be naive to me. (I can add that the signs have made a different to me in at least some ways. When I see that, say, a not very tasty scone at Starbucks has 300 cal., at the least it makes me think I should eat something I'd like more if it's going to have that many calories.)

Posted by: Matt | Apr 3, 2014 9:44:38 AM

Post a comment