Wharton's Benjamin Lockwood discusses his research on determining the optimal rate for so-called sin taxes, like Philadelphia's tax on soda.

It’s been two and a half years since Philadelphia became the first large American city to charge a soda tax. The tax rate of 1.5 cents per ounce — 30 cents on a 20-ounce bottle or $1 on a 2-liter bottle — was designed to improve health outcomes by lowering consumption of sugary drinks. The tax has raised $200 million in revenue earmarked for pre-K education and other city programs, but the policy has met with strong resistance by soda manufacturers, grocery store owners and restaurateurs who say the tax has hurt their businesses. In a recent study, a team of economists analyzed the results of the so-called sin tax on soda to determine the optimal rate. “Regressive Sin Taxes, with an Application to the Optimal Soda Tax,” is the title of the paper by Benjamin Lockwood, Wharton professor of business economics and public policy, Hunt Allcott, economics professor at New York University, and Dmitry Taubinsky, economics professor at the University of California Berkeley. Lockwood spoke during the Knowledge at Wharton radio show on SiriusXM about the research and what it means for other cities considering a similar tax. (Listen to the podcast at the top of this page.)

An edited transcript of the conversation follows.

Knowledge at Wharton: Can you tell us broadly about the research?

Benjamin Lockwood: To start off, we’re thinking big picture. We know that there are some benefits of these kinds of taxes, that you can reduce consumption of harmful goods, you can reduce social costs for health care spending. We also think there are some costs, some downsides of these kinds of taxes. Fundamentally, people like doing things that are sometimes not great for them, and we want to take that into account. There’s also some concern that these kinds of taxes can be regressive because poorer consumers often buy soda at higher rates. We wanted to take all of these different things into account and figure out, on net, are these taxes beneficial for society or are they harming society?

Knowledge at Wharton: What did you find out?

Lockwood: On net, our estimates suggest these taxes are beneficial. There are some substantial externalized health costs, meaning there are some costs that the rest of society bears every time someone drinks soda. It’s something like 10 cents for every 12-ounce can of Coke, so not a trivial amount. Our estimates also suggest there are costs that consumers don’t take into account when they’re drinking the soda, so they’re a little better off if they reduce their consumption a bit. Those two effects actually outweigh the concerns about regressivity and the benefit that people get, the sort of enjoyment you get, from drinking a Coke. On net, something like a tax between 1 cent and 2 cents per ounce is probably a good idea.

Knowledge at Wharton: How did you arrive at that number?

Lockwood: It comes from a few different components. There is an estimate that there is a kind of negative external social cost from drinking sugary beverages. When you drink a soda, you’re more likely to get diabetes, you’re more likely to be obese. That’s a cost that Medicare has to bear. That’s a cost that all the rest of us have to bear. That’s estimated to be a little shy of 1 cent per ounce of soda consumption, so that right there is a decent-sized cost.

Then there’s also the internal cost, the cost that people don’t think about but that they will bear when they’re drinking soda. These things are hard to estimate, but I think a couple observations are helpful. First, among soda drinkers, we find that more than half of them say they should cut back. That already suggests that they would probably be better off if they drank less. We can use that kind of information to infer what those internal costs from overconsumption are, and we get an estimate of about another 1 cent per ounce. So, that adds up to almost 2 cents per ounce. Because poorer consumers pay a little more of it, it’s good to reduce it a little below that. That’s where we get this estimate of 1 cent to 2 cents per ounce.

Knowledge at Wharton: If other cities are considering enacting a soda tax like Philadelphia has done, is that the ballpark rate?

Lockwood: Our guess is that the optimal tax is somewhere between that 1 cent to 2 cents per ounce range. Now, there are a couple caveats to that. One is that we were mostly focusing on the optimal tax at the state level or even the national level. When you drill down to really small localities, there’s some concern that people can go outside the city or across the border to shop tax free. There is some evidence that that happens to some extent in Philadelphia and a few other places that have passed these kinds of taxes, so that undermines their effectiveness a little bit. They’re probably still a good idea, but they’re not quite as beneficial as they would be if people didn’t do that cross-border shopping.

“We want to allow for the possibility that people do like drinking soda and sometimes do like doing things that are a little bit bad for them.”

This kind of cross-border shopping is more of a concern in really small areas and in areas where it’s easy to go across a border and purchase elsewhere. In larger cities or cities that are separated from nearby untaxed areas, it’s a bit less of a concern. The estimates in Philadelphia are that, even after you account for this kind of cross-border shopping, overall soda consumption is still down by over a third because of this tax.

Knowledge at Wharton: The idea is to try to reduce the consumption as much as you possibly can, but is there a specific goal like 50% or 75%?

Lockwood: This is one of the areas where I think economists sometimes think about things a bit differently from the general public. We want to allow for the possibility that people do like drinking soda and sometimes do like doing things that are a little bit bad for them. There are some health costs of rock climbing, and we don’t want to prohibit rock climbing. Our goal is to think about what tax will get people to act as if they were fully cognizant of the long-term health costs, even if that does mean they would still have a soda now and then. We think that’s fine. We just don’t want people consuming so much that it suggests they don’t understand what the long-term health effects are, and similarly for the health costs that fall on the rest of society.

Knowledge at Wharton: Can you talk about the idea of this being called a sin tax and what effect that specific terminology has on the purchasing process?

Lockwood: Let me first say what we mean by the technical definition of sin tax and then a bit about what the implications of that language are. There are a number of different reasons why you might want to have taxes of any sort in place. You might want to raise revenues for government projects, for public goods and the like. You might want to reduce inequality by taxing the rich more and redirecting those resources toward the working class. Or you might want to try to change people’s behaviors, and you can do that through taxes or through subsidizing things that you think are good activities. When we talk about sin taxes, we’re really talking about that last category of taxes where you’re trying to change people’s behavior, both because that behavior might have a cost for the rest of society and because we think there are health harms that people aren’t fully aware of when they are making their decisions.

One interesting side effect of these kinds of taxes is that sometimes the mere discussion of them creates extra awareness about the costs and the downsides of consumption. I think that’s a little less relevant for soda. Many people have the sense that that’s not a great thing for you. But there are many things like fruit juices and Gatorade that are full of sugar and are probably not great either, but people think of them as healthy consumption.

Knowledge at Wharton: There is also a significant difference when you’re talking about a tax on soda in comparison to a tax on cigarettes. Cigarettes have been proven to have an impact on health, and that warning is put on the labels. So, I think there’s a difference in the perception.

Lockwood: Yes, although I would emphasize that when cigarette taxes were first passed, there was still a lot more debate from tobacco companies and the like about whether these taxes were creating long-term health benefits. They would often say there is no proven health benefit because the health benefits take a long time to see as a result of the tax. Right now, you’re seeing similar sorts of arguments with respect to these soda taxes. I think that the public health research on the effects of sugar is fairly clear at this point. Because the taxes themselves are pretty early, it will be a few years before we really see the benefits from those.

Knowledge at Wharton: What did your research find about the effect of a soda tax on people with lower incomes?

“When you impose a tax on soda, people do indeed reduce their soda consumption. In fact, they reduce it by so much that their overall spending on soda goes down.”

Lockwood: This is one of the things that we most wanted to understand with this research. Are these kinds of taxes making lower-income classes worse off? You hear very different arguments in this area. U.S. Sen. Bernie Sanders originally opposed the soda tax on the grounds that it would take money away from poor consumers in Philadelphia. You’ll also sometimes hear people make the argument that because these taxes reduce health harms like diabetes and obesity, which are most prevalent among poorer households, they create a more progressive benefit because those households won’t suffer these sorts of long-run harms. We take both of those things into account in our estimates. Indeed, you see both of them are concentrated on poorer households. They do pay more through the tax, but at the same time, they do get the resulting health benefits. On net, it looks like the tax basically has a flat profile of benefits across the income distribution.

Knowledge at Wharton: Part of what has been discussed in Philadelphia is the allocation of the tax revenue. I mentioned pre-K education. Mayor Jim Kenney has talked about using the money to get other services back up and running, like public pools. He views that as a benefit for people in lower income brackets. What do you think?

Lockwood: Yes, I think that’s probably right. This has been an interesting change in the way that people talk about these sorts of taxes. They’re emphasizing the way that the revenue is going to be used and how it can often be targeted back in some poor communities that end up paying more of the tax.

Knowledge at Wharton: Why tax diet soda?

Lockwood: This is a great question. We have a number of recommended principles for policymakers when they’re thinking about designing these taxes. One is that they think about not including diet soda among the taxed goods. Our sense, and the sense from public health researchers, is that there may be some health downsides of artificial sweeteners, but they do not seem to be as large or as well-estimated as the negative costs of sugar. Although in the best possible world, you’d like to have people switching to water or something, it’s probably on net better to have people switch to diet and away from regular. When you tax diet goods too, it shuts down that channel of substitution.

In Philadelphia, diet soda is taxed. In many other cities that have imposed these kinds of taxes, diet soda would be exempt. Our best guess at this point is that it probably is a good idea to exempt diet sodas or at least to have a much lower tax on them. That’s one way in which Philadelphia’s tax could probably be improved a little bit.

Knowledge at Wharton: The argument made by companies not in support of the tax, such as restaurants and grocery stores, is that lower consumption costs jobs. What do you think of that stance?

Lockwood: This is an interesting argument. There is some new data out from Philadelphia, and some researchers have been looking at this. There is no evidence of a net job reduction from these sorts of taxes. But one thing that I want to emphasize is that when you impose a tax on soda, people do indeed reduce their soda consumption. In fact, they reduce it by so much that their overall spending on soda goes down, rather than up.

It’s not as if that money has disappeared. What that means is that they have more money in their pocket to spend on other things. That spending also generates jobs in other areas. The pre-K spending generates jobs for additional teachers. What I think we want to think of at a macro level is that the money is still there and being spent in different ways, it just might result in a reallocation away from having quite so many supermarket aisles full of sugary beverages and toward other things.

Knowledge at Wharton: How much further can we take this research in terms of the health side? The economic side is important, but the health side continues to be the main focus. It’s a message that won’t always be followed.

Lockwood: I think that probably the best basket of policies involves a number of different things to try to get those messages across: taxes, health education, better labeling campaigns to make it clear which things have lots of sugar and which things don’t, and better education about this topic in general.

Knowledge at Wharton: Do you think there is any appetite for a federally imposed soda tax?

Lockwood: I would be surprised to see a federal tax on soda anytime soon. The reason we think about that kind of policy in the paper is to get away from these issues of cross-border shopping, which would make it a little more complicated. It would not surprise me if, in the next five years or so, we do see some discussion of sweetened beverage taxes at the state level. There’s already been some discussion of that in California and I think some early stage discussion potentially in Massachusetts.

“I do think as we build the case and show the economic benefits of these kinds of taxes, more of them will be adopted, as was the case with cigarette taxes.”

Knowledge at Wharton: Can your research apply to other sin taxes that are in place right now?

Lockwood: Absolutely. I think we can better understand things like the regressivity of cigarette taxes this way, thinking about both the costs to core consumers but also the health benefits that they experience. If you flip the sign on a tax and interpret it instead as a negative subsidy, we can think about energy efficiency subsidies. It’s when you subsidize the purchase of electric cars, for instance, but a lot of those subsidies go toward richer consumers, which is sort of a regressive aspect we might not love. You can use this sort of insight for those questions, too.

Knowledge at Wharton: Would you agree that the challenge is making the word “tax” palatable and driving home the benefits?

Lockwood: Yes, absolutely. That is one reason I think cities have found it helpful to really tie tightly together the tax and the uses for the revenues when they’re making the case to the public, because then people can see tangibly what this money is going toward and how they’re being made better off by it.

Knowledge at Wharton: Does median income play a role in having sin taxes in a city?

Lockwood: In principle, it could. That’s one reason we were really interested in how different parts of the income distribution would be made better or worse off by these kinds of taxes. Because you have this balancing out of the amount of money that people pay with the amount of health benefits that they get, we find that to be pretty flat across the income distribution. So, it does not suggest that it’s much worse to have these taxes in areas that are more poor or more rich or whatever. It’s sort of beneficial across the board.

Knowledge at Wharton: What are the next steps this your research?

Lockwood: It’s sometimes difficult to convince someone who’s firmly opposed to these kinds of taxes to implement one in the first place. But there are many places where these kinds of policies are already being discussed, and there is some opportunity to tweak it a little bit to make the tax a bit better. One thing that we’re focusing on right now is trying to better understand the improvements that could be made. We’re looking at encouraging places to switch from volume-based taxes, where it’s 1 cent to 2 cents per ounce, to a sugar-based tax, where it’s some number of cents per gram of sugar, so that people are encouraged to slightly lower sugary drinks even if they’re going to consume a sugary drink either way. That’s an example of another ongoing project that we have that we think could improve these kinds of policies.

Knowledge at Wharton: Soda companies are fighting this tooth and nail. But by bringing out more and more of these products that have low sugar or no sugar, they seem to understand that this is an issue.

Lockwood: And that’s one of the goals of these kinds of taxes. Again, if you do have a sugar-based tax or if you exempt diet drinks, you can have an impact on the way that these products are manufactured and marketed.

Knowledge at Wharton: Do you expect to see more places enact soda and other kinds of sin taxes as more research shows the benefits?

Lockwood: That’s my guess. There are always bumps in the road, like in the case of Chicago’s repealed soda tax. But I am an optimist for research and for evidence over the long run. I do think that as we build the case and show the economic benefits of these kinds of taxes, more of them will be adopted, as was the case with cigarette taxes.