Saturday, August 27, 2016

Berkeley "soda tax" reduced sugar-sweetened beverage consumption and increased water consumption

In the American Journal of Public Health this month, Jennifer Falbe and colleagues found that the penny-per-ounce Berkeley soda tax succeeded in reducing sugar-sweetened beverage (SSB) consumption.

The study asked respondents about soda intake, in Berkeley and in comparison cities of northern California (Oakland and San Francisco, which did not have a new tax), before and after the new Berkeley tax was implemented in March 2015. The findings were remarkable. SSB consumption fell 21% in Berkeley and rose 4% in the comparison cities during the same period.

The SSBs include caloric soda, of course, but also some kinds of other sweetened drinks. However, for various reasons (including sound nutrition reasons plus perhaps political reasons), the tax did not affect milk drinks or 100% fruit juice. Therefore, it is important to understand what other beverages people substituted for soda. The study does not answer all my questions on this point, but it did find that water consumption increased in Berkeley at the same time that SSB consumption fell. Water consumption increased 63% in Berkeley, significantly more than the increase in the comparison cities during the same period. This was reassuring.

A good way to standardize estimates of tax effects is to report an "elasticity" -- the percentage change in consumption for each 1% change in price. A typical elasticity estimate for soda is about -1.2, meaning that the price increase in Berkeley (about 8%) would have been expected to generate a consumption decline of about 10%. The authors took care to confirm that the estimated consumption decline of 21% was significantly different from zero, which is the standard statistical way of making sure the estimates were not a random statistical fluke, but they cannot really be sure the true impact is exactly 21% rather than 10%. They sensibly discussed the possibility that "early reaction to the tax ... could rebound and settle closer to a 10% reduction in consumption."

Even if the impact were a 10% reduction, this study has important public health implications, providing I think the strongest evidence so far that a tax would reduce SSB consumption.

I encourage my colleagues in agricultural and applied economics to read this study. There is a long tradition in my profession of doubting the potential impact of such taxes. In the Washington Post in 2015, Tamar Haspel quoted University of Minnesota applied economist Marc Bellemare saying the results at that time were "not robust." Haspel also quoted my Friedman School colleague and friend Sean Cash saying that product formulation, rather than taxes, are the way to go: "If we could achieve a 5 percent reduction by reformulation, that would swamp what we can achieve with consumer-level intervention.” The TuftsNOW site quoted Sean casting further doubt on taxes: "All studies suggest that for food in general, we’re not particularly responsive to price." Oklahoma State University economist Jayson Lusk, who also is president of the Agricultural and Applied Economics Association (AAEA), has blogged several times about soda taxes, agreeing with most of the Tamar Haspel column  in the Washington Post, and concluding stridently: "I'm sorry, but if my choice is between nothing and a policy that is paternalistic, regressive, will create economic distortions and deadweight loss, and is unlikely to have any significant effects on public health, I choose nothing" (emphasis added).

In the Salt this week, NPR reporter Dan Charles quotes Berkeley researcher Kristine Madsen on whether the new estimates of SSB reduction are large enough to matter for public health. "Madsen says a 20 percent reduction in consumption of sugar-sweetened beverages would be enough to reduce rates of obesity and Type 2 diabetes in years to come. 'This would have a huge public health impact if it were sustained,' she says." I think most experts in public health nutrition would agree with Madsen's assessment.

This week, Jayson's blog post on the Berkeley study raises some measurement issues, but recognizes that these issues are unlikely to overturn the main result. He writes, "All that said, I'm more than willing to accept the finding that the Berkeley city soda tax caused soda consumption to fall. The much more difficult question is: are Berkeley residents better off?" This is a question that surely will be discussed heavily in the next couple years as more municipalities experiment with such policies.

10 comments:

Timberati said...

The purpose of a Pigou tax such as Berkeley's soda tax, as I understand it, is to "pay forward" the damage of the externality. In this instance, it would be the cost of accommodating larger people (obese people live shorter lives, so the medical costs to society are actually lower). As Tim Worstall points out (on a Pigou tax for carbon), "In economic theory, the problem here is that my actions that create emissions also damage someone else. But I don't have to pay for the damage I've caused. This is called an externality and the economists' solution is something called a Pigou Tax. That is, we add a tax equal to the damage I'm doing, so that I do pay for that damage."

I suspect, that is not what the citizens of Berkeley thought they were doing.

Link to Worstall post: http://blogs.telegraph.co.uk/finance/timworstall/100016929/why-the-answer-is-a-carbon-tax/

Parke Wilde said...

Thanks for the comment, Timberati.

First, there is a large literature in economics on the externality component of poor nutrition. Economists tend to think it reasonable for public policy to address the harm that poor nutrition people does to other people (for example, through premium prices in insurance risk pools and the cost of government medical programs). Here is just one example: http://www.nber.org/papers/w11529

Second, other disciplines, such as public health nutrition, implicitly treat reducing the total health burden as a legitimate goal of public policy.

I couldn't quite tell from your comment which of these perspectives you were taking. In both cases, I can understand why policy-makers contemplate using tax policy to influence soda consumption.

Timberati said...

Thank you for the clarifications, Parke.

I tend to be a libertarian who likes economics. I think we should have the liberty to make stupid life choices up to a point. That point begins when my liberty interferes with your.

This may help to illustrate my bent: I favor mandatory vaccines due to harm non-vaccinated persons can do to those for which a vaccine is not effective (herd immunity). I'm simply not sure the obesity issue is in the same category as vaccines, and actually there are numbers from the UK (Worstall) that show that fat people, smokers, and heavy drinkers cost the health system less because they die younger.

- Norm

Worstall's obesity, smoking, and alcohol article: http://www.forbes.com/sites/timworstall/2012/03/22/alcohol-obesity-and-smoking-do-not-cost-health-care-systems-money/#743fac3313c2

Timberati said...

If you want to find typos, just hit the "send" button. Your = yours.

Parke Wilde said...

Timberati, that Worstall opinion piece is a fat-shaming stew of cruelty, with little of interest to say about the merit of policies such as the Berkeley tax. Worstall wishes "lardbuckets" an early death so that they can save society the cost of their medical care.

Sage said...

That is not the only purpose or use of a Pigovian tax.

Tamar Haspel said...

Hey Parke -- I have to weigh in here, because I think you have misrepresented my position in two ways. For starters, you cite Sean's quote about reformulation, but leave out the part where I'm suggesting (as he is) that what would drive reformulation is a tax on sugar. Since you leave that Your piece implies that we're relying on some kind of industry do-gooding to make that happen. Far from "doubting the potential impact of such taxes," that's actually suggesting that one of those taxes -- on sugar, in the supply chain -- would have significant impact.

The other reason is that, by omitting my position on taxes, but citing Jayson's at some length, in the same sentence where you say he agrees with most of my column, you imply that I also stridently oppose soda taxes. To do my position justice, I think you should add that we certainly disagree on that.

My position on soda taxes is that they're likely to reduce consumption, at least a little, but less likely to have a public health benefit, as that benefit depends entirely on what people do instead (which we have no idea about). I am nevertheless in favor of them, because fighting obesity requires resources, and getting those resources from a tax on a product that clearly contributes to the problem seems reasonable.

Parke Wilde said...

Hi Tamar,

Thanks for sharing your thoughts in this comment thread.

I did not misrepresent your position -- not in "two ways" and not even in one way.

My blog post is about SSB taxes similar to the Berkeley tax.

I did not summarize your policy position on soda taxes, not even to drop a hint or insinuation about it. I have a "live and let live" attitude about any policy position you may take, and I have no interest in reporting those policy positions in my blog.

I also did not summarize your views on other types of tax, at other stages of the supply chain, such as a tax on sugar as an ingredient. I had good reason to focus my blog post on taxes similar to the Berkeley tax, and that is my prerogative as an author.

What I did was correctly and fairly report your 2015 summary of evidence on the likely effect of a Berkeley-style tax on (a) SSB consumption and (b) public health.

Regarding (a), you favorably quoted experts who said the likely impact of such a tax on consumption would be small. Despite the new evidence from the AJPH, in your whole tweetstorm and in your comment here, you have not revised that assessment upwards, as the evidence merits.

Regarding (b), you say that you doubt the public health nutrition or obesity impact of a Berkeley-style tax. By contrast, I think that a 10% consumption change, and certainly a 21% consumption change, would be sufficient to have a public health impact. While you may have "no idea about" what else people consume, the Berkeley study did offer evidence on substitution of water in place of SSBs.

Feel free to agree or disagree on either or both points. I am happy to learn from your own reading of the evidence. If you refrain from false accusations that I have misrepresented your position, that conversation will have more light and less heat.

Tamar Haspel said...

Parke, since my contention is that the misrepresentation came from omission, and you acknowledge that you omitted what wasn't relevant to your thoughts on the Berkeley soda tax (i.e. the function of a supply-chain tax, and my policy position), I certainly see why we disagree. If you simply don't care what readers conclude about my position, as it is irrelevant to your post and not your job to elucidate, we just have conflicting priorities. I do care what readers conclude about my position.

The issue at hand, though, is the study about the Berkeley tax, and I think it shows some promise that soda taxes may be effective -- which is why I tweeted it when it first came out. I also, however, think that using one study of self-reported consumption to revise a conclusion based on many countries' experience with soda taxes (gleaned from sales data) is a bit premature. There has been a great deal of anti-soda messaging in the US, and that may give a tax more oomph than it might have had, absent that messaging. But, by the same token, the messaging may bias self-reported consumption data, so it's a double-edged sword.

Parke Wilde said...

Okay, I'll buy that summary of our tussle about misrepresentation.

And I take seriously your warning that future research may not corroborate the Berkeley study.

Thanks for sharing your thoughts here!