Research on The (lack of) Effectiveness of Bloomberg's Large Soda Ban

Much has been written about the merits or demerits of Bloomberg's large soda ban (here was my recent take on it in the New York Daily News).​

However, there has been much less actual research conducted to determine whether such restrictions might curb consumption or on how retailers might respond.  Well, some researchers from UC San Diego conducted a small scale study on the issue that was just published in the journal PLoS One.

What they showed is that food companies can get around the ban by offering bundles of smaller-sized drinks and that people respond in kind by buying more soda!  The study reminds me of what happened when San Francisco tried to ban giving away toys in Happy Meals; McDonalds decided to instead sell them for a very low price ($0.10).  

That's the problem with a lot of these regulations - people and companies find a way around them in ways that the regulator couldn't envision and, as this PLos ONE study shows, it might even lead to weight gains.  It's like squeezing a balloon - the air doesn't leave it just moves to a different place.  Banning large soda or Happy Meal toys doesn't diminish demand for these items, it just causes people to seek out alternative means to get them.  ​

Here is the study abstract:​

Objectives

We examined whether a sugary drink limit would still be effective if larger-sized drinks were converted into bundles of smaller-sized drinks.

Methods

In a behavioral simulation, participants were offered varying food and drink menus. One menu offered 16 oz, 24 oz, or 32 oz drinks for sale. A second menu offered 16 oz drinks, a bundle of two 12 oz drinks, or a bundle of two 16 oz drinks. A third menu offered only 16 oz drinks for sale. The method involved repeated elicitation of choices, and the instructions did not mention a limit on drink size.

Results

Participants bought significantly more ounces of soda with bundles than with varying-sized drinks. Total business revenue was also higher when bundles rather than only small-sized drinks were sold.

Conclusions

Our research suggests that businesses have a strong incentive to offer bundles of soda when drink size is limited. Restricting larger-sized drinks may have the unintended consequence of increasing soda consumption rather than decreasing it.

While the study findings are intriguing, it must be said that the study is far from perfect.  For example, the study involves a bunch of college students making a number of hypothetical choices.  I'd much prefer to see an experiment where people actually had to pay (and eat) what they bought.  Moreover, as the study authors readily acknowledge, the study doesn't reveal whether people would actually drink both sodas or just give one to a friend, nor did it differentiate between diet or full calorie soda.  Thus, there appears to be fertile ground for additional research. 

Do Scientists Mislead?

"Yes" seems to be the unfortunate conclusion that stems from this paper by Mark Cope and David Allison published in the International Journal of Obesity.  ​Scientists may not (or may) distort their own research findings, but Cope and Allison show, pretty convulsively, that there is a general pattern of distorting the findings of others.   

They attribute this to a "white hat bias":​

which we define to be bias leading to distortion of research-based information in the service of what may be perceived as righteous ends.

Cope and Allison found two studies related to soda consumption that:

had both statistically and non-statistically significant results on body-weight, body mass index (BMI) or overweight/obesity status which allowed future writers to potentially choose which results to cite, and were also widely cited, permitting quantitative analysis of citations.

​Then, they looked at how other scientists subsequently cited the findings in their published papers.  Did they focus on the negative findings (that soda doesn't affect weight, etc.) or the positive findings (that soda does affect weight, etc.): ​

The majority, 84.3% for [2] and 66.7% for [3], described results in a misleadingly positive way to varying degrees (i.e., exaggerating the strength of the evidence that [nutritively-sweetened beverage] reduction showed beneficial effects on obesity outcomes). Some were blatantly factually incorrect in their misleading statements, describing the result as showing an effect for a continuous obesity outcome whereas no statistically significant effect for continuous obesity outcomes was observed. In contrast, only four papers (3.5%) were negatively misleading (i.e., underplayed the strength of evidence) for [2] and none were negatively misleading for [3]. Only 12.7% and 33% of the papers accurately described complete overall findings related to obesity outcomes from [2] and [3], respectively.

​

They went on to document a similar pattern for studies on the effects of breastfeeding. They ​conclude:

Evidence presented herein suggests that at least one thing has been demonized ([nutritively-sweetened beverage] consumption) and another sanctified (Breastfeeding), leading to bias in the presentation of research literature to other scientists and to the public at large, a bias sufficient to misguide readers. Interestingly, while many papers point out what appear to be biases resulting from industry funding, we have identified here, perhaps for the first time, clear evidence that white-hat biases can also exist in opposition to industry interests.

Fat taxes may be even less effective than previously thought

An article that just appeared in the most recent issue of the American Journal of Agricultural Economics by Yuqing Zheng of RTI and Edward McLaughlin and Harry Kaiser of Cornell University presents some interesting thoughts regarding fat and soda taxes.

Most of the studies on fat and soda taxes use elasticities of demand to simulate the effectiveness of a tax.  The elasticity of demand tells us how responsive consumption of a product is to a change in it's price.  So, for example, an elasticity of demand of -0.6 would tell us that for every 1% increase in price, consumption would fall by 0.6%

Zheng and colleagues point out, however, that most people don't "see" the tax when they're shopping.  It only shows up on the bill when you check out.  As such, it is incorrect to directly use the price elasticity of demand  ​to infer how consumption of soda or fatty foods will change after a tax.  

They show that an excise tax (​which is levied on the supplier rather than the consumer) is probably more effective at reducing soda (or fat) consumption than a retail sales tax, but even excise taxes are likely to to be less effective at reducing consumption than an equivalent price increase.  

As a result, researchers need to adjust the demand elasticities before calculating the effects of a soda or fat tax (regardless of whether the tax is excise or retail).  Yet, almost no one does this.  The researchers show, however, that the required adjustment can be fairly dramatic.  In their "baseline" scenario, they show that the elasticity of demand needs to be reduced by a factor of 0.14.  So, if the elasticity of demand was -0.6, then we'd project a 1% increase in sales tax would only reduce consumption by 0.6*0.14 = 0.084%.  Instead, previous research on this topic has applied the original elasticity (e.g., 0.6) rather than the adjusted elasticity (e.g., 0.084).  Clearly, the adjusted elasticity will imply much lower effectiveness of the tax.

The authors conclude:​

Therefore, even a large increase in the sales tax rate on food and beverages will only reduce demand by a moderate degree.

and

our analysis of sales and excise taxes offers some explanation (e.g., imperfect tax knowledge, slow learning) on why the impact of sales tax is so difficult to detect, thus bridging the gap between the simulation studies and the empirical findings

Robert Frank on Soda Taxes

A couple days ago, the Cornell economist Robert Frank weighed in at the New York Times on Bloomberg's soda ban.  

The first part of the article is mainly about negative vs. positive freedoms.  Frank wants the government to guarantee ​positive freedoms (how many, I'm not sure) whereas I'd prefer the government primarily focused on guaranteeing negative liberties.  This is an old debate (I'm reminded of the chapter in Hayek's Road to Serfdom where he talked about the successful efforts to redefine the word "freedom.") and although I disagree with Frank, I at least understand where he is coming from (aside from the ad hominem connections between soda and tobacco).

What confuses me (because he's an economist) are Frank's economic arguments for a soda tax.  ​

In the 2nd to last paragraph he says:​

The case for reintroducing such a proposal is strong. We have to tax something, after all, and taxing soft drinks would let us reduce taxes now imposed on manifestly useful activities. At the federal level, for example, a tax on soda would permit a reduction in the payroll tax, which would encourage businesses to hire more workers.

​Yes, if we want public roads and schools, we do need to tax something.  But, I would have thought than an economist would argue that we'd want to tax things that create the least deadweight loss. And I would have thought there would be mention that we'd want taxes that are the least regressive (i.e., where the burden is not most heavily paid by the poor).

Maybe Frank is right and we should reduce payroll taxes, but how do we know soda taxes are the solution and are any less distortionary or are more progressive?  And why would soda taxes be the next best alternative to replace payroll taxes from an economic point of view?  Elsewhere, Frank has (very reasonably in my opinion) argued for across-the board consumption taxes in lieu of income taxes; so why now a soda-tax rather than an across-the-board consumption tax? There is no attempt to answer these important questions.     

Yet, he goes on in the concluding paragraph:​

Evidence suggests that the current high volume of soft-drink consumption has generated enormous social costs. So to those who have lobbied successfully against a soda tax, I pose a simple question: How do the benefits of your right to drink tax-free sodas outweigh the substantial costs of defending it?

​I disagree.  We don't know that soft-drink consumption per se has enormous social costs.  We have epidemiological-type studies/simulations based on correlations and assumptions but we don't know much from randomized controlled trials.  We also don't know that the costs (whatever they might be) are social rather than private.  As is well articulated in this article by Jay Bhattacharya and Neeraj Sood published in the Journal of Economic Perspectives when it comes to obesity more generally, the costs are primarily private​. Moreover, the expenses to Medicare and Medicare are really just transfers not deadweight loss.

Frank ends with a question: ​"How do the benefits of your right to drink tax-free sodas outweigh the substantial costs of defending it?"

First, I don't think most opponents of soda taxes would defend "tax free" but rather argue against special exceptions/taxes for one type of food vs another (because, among other reasons, the definition if "soda" isn't all that clear - and there are many substitutes for sodas that may be more calorie dense).  But, more generally, the answer is that the economic research shows soda taxes to have small effects on consumption, yet with large deadweight loss, and no meaningful impact on obesity.  The answer, in short, is that I defend my perspective using economic logic and the empirical results from economic studies.  I'm willing to be persuaded by new evidence or economic reasoning, but Frank didn't offer anything new.  I'm a bit surprised an economist of the caliber of Frank didn't meaningfully engage with economic evidence we do have.   

Does Sugar Cause Death? Or Increase Life?

All over the news today I've read about some research that was presented by some Harvard professors purporting to show that sugary beverages lead to 180,000 deaths worldwide and 25,000 deaths in the US (here is Time, ABC,  ​Forbes, or a press release by the American Heart Association).

​I haven't yet been able to get my hands on a copy of the actual paper (which isn't yet published) but from all accounts, this is not hypothesis-testing-science sort of research but rather assumption-driven-simulation sort of research.  I've done some of that kind of work myself and it can produce some useful insights, but the results are only as good as the assumptions, and I suspect (though don't know) the simulation is assuming causality where there is only correlation.

While we are looking at correlations, we might as well have some fun with it.  About two weeks ago, I blogged about (and reanalyzed data from) some research by anti-sugar crusader Robert Lustig that was published in PLoS ONE.  In short, they showed a positive correlation between countries that eat a lot of sugar and prevalence of diabetes in the country.  The correlation was widely interpreted as proving a causal link between sugar and diabetes.  So, let's apply that same logic to the topic of death and sugar consumption.

I took the same sugar consumption data as I talked about in my previous post​.   Then, I merged it with life expectancy by country as compiled by the UN (you can download the data from Wikipedia).  Here is the relationship between average life expectancy in a country (in years) and per capita availability of sugar (in kcal).

We've clearly got a positive correlation here.  In fact if I run a population-weighted regression, I find that for every 100 kcal increase in sugar availability, life expectancy increases 1.9 years!  The p-value is even less than 0.001!​  Drink more sugar and you'll live longer!

Now, do I actually think increased sugar consumption increases life expectancy?  Probably not.  Richer countries can probably afford to buy lots of things (like more sugar) and richer countries are also likely to have higher life expectancies (because they can also buy more health care).  But, it does make you wonder what's going on if the Harvard researchers are right and all that sugar is causing so many deaths. ​

I rather suspect, a lot of deaths due to other problems are being attributed to sugar from sodas.  I'm reminded of this passage from Eric Oliver's book Fat Politics as it relates to attributing deaths (or medical costs) to obesity.

The researchers who estimated that obesity costs us 100 million dollars a year did so by calculating all the expenses associated with treating type 2 diabetes, coronary heart disease, hypertension, gallbladder disease, and cancer . . . Once again, they simply assumed that if you got heart disease or breast-cancer it was because you were fat.