Wharton's Benjamin Lockwood discusses his research on how 'sin taxes' affect consumer behavior

‘Sin tax’ is defined as a tax on a product that can be harmful to a person, such as cigarettes or sugary drinks. In many cases, these taxes are an incentive to lower consumption and improve health. But sin taxes can disproportionately hurt lower-income consumers, while wealthy shoppers enjoy tax breaks on items only they can afford, such as energy-efficient windows and appliances. A recent study by Benjamin Lockwood, a Wharton professor of business economics and public policy, and coauthor Dmitry Taubinsky from Dartmouth College examines the impact of sin taxes and whether there is a middle ground. The researchers also look at what is being called “revenue recycling,” where these taxes can be used to fund initiatives that benefit lower-income consumers. Lockwood recently spoke about their research on the Knowledge at Wharton show on Wharton Business Radio on SiriusXM channel 111. (Listen to the podcast at the top of this page.)

An edited transcript of the conversation follows.

Knowledge at Wharton: Sin taxes are an important topic here in Philadelphia, where there is a tax on sugary drinks.

Benjamin Lockwood: Absolutely. It’s been happening in Philadelphia, and we’ve also seen these implemented in Chicago, San Francisco, Berkeley, Oakland and Boulder, Colorado. There’s a growing policy wave in favor of these kinds of policies, so it seems like a good time to be looking at it and trying to understand some of these implications.

Knowledge at Wharton: What did you find in your research?

Lockwood: The way that economists generally think about these kinds of taxes is that sugary beverages have health consequences. They can give rise to things like diabetes or stroke or heart disease, and these are pretty big in magnitude. A couple of years ago, there was a study estimating that if people were to reduce their sugary beverage consumption by around 20%, then the health benefits that they would reap would be something akin to giving them each a check between $100 or $300 each year. These are pretty big numbers.

On the other hand, from an economist’s perspective, it’s not enough for something to have negative consequences to justify taxing it. Things like rock climbing have negative health consequences, potentially. Driving a car has lots of negative health consequences. The key question from an economic policy perspective is whether people are taking into account these negative effects when they’re making their consumption decisions. In particular, a reason for concern about sugary beverages is that often the negative health consequences come a long time after the date of consumption. You get diabetes or heart disease much later in life.

There’s a growing literature in behavioral economics that studies the tendency for people to underweigh distant consequences and overweigh the upfront benefits or costs of doing something. This can explain everything from why we save less for retirement than we should or intend to, or why we exercise less than we ought to. A reason for being interested in sugary soda and sugary beverages is that those [choices] also have this kind of discrepancy between the upfront joy of sipping a soda and this delayed health consequence that happens far down the road.

“From an economist’s perspective, it’s not enough for something to have negative consequences to justify taxing it.”

Knowledge at Wharton: In a lot of urban areas, people financially may not have another option in terms of drinking a soda, compared with drinking bottled water. It becomes a life issue that a lot of these people are not able to overcome.

Lockwood: Right. Part of what you’re bringing up here is the question of what people can afford and how these kinds of taxes hit poorer consumers versus richer consumers. This question is the fundamental one of our research. There have been studies of how these kinds of taxes can have beneficial health consequences by reducing consumption. But there is also concern about an unintended side effect of that kind of policy — that it tends to fall heavily on poorer consumers. We know that poorer consumers tend to consume things like cigarettes and soda at higher frequencies than richer consumers do. Survey evidence suggests that at the bottom of the income distribution, people drink about twice as much sugary soda than at the top of the income distribution.

This paper looks at the regressivity consequences of these kinds of taxes and tries to get a handle on them. How do we weigh those consequences against the potential health benefits from imposing these kinds of taxes?

Knowledge at Wharton: In your estimation, are sin taxes a good thing for the consumer in general?

Lockwood: This is the million-dollar question. What is the overall impact? And if we should have a soda tax, how big should it be? Philadelphia’s soda tax is 1.5 cents per ounce. I believe Boulder’s is 2 cents per ounce. Most of the others have been 1 cent per ounce. As cities go forward trying to weigh these policies, there is this question of what the magnitude should be and whether we should have this kind of tax at all.

The key thing that we explore in our paper is that what matters for these regressivity costs is how much people respond to these taxes when they are imposed. There’s often an initial intuition that these taxes must be really bad for poor consumers because then they have to pay more out of pocket. That’s exactly right, if people don’t respond to the tax at all — if they don’t reduce their consumption. Of course, poor people end up paying more.

On the other hand, if people end up reducing their consumption a lot in response to the tax, then things get a lot trickier and a lot more interesting. The people who get the greatest health benefits from that reduction are the people who were consuming the most sugar to begin with, which tends to be poorer consumers. So, if people are responding a lot to the tax, then these kinds of regressivity costs are actually a lot smaller. In fact, some of the health benefits can be really concentrated on poor consumers, which is something that the government is interested in.

To answer the question you raised — how should these taxes exist and how big should they be — the key question is how much people reduce consumption in response to the tax. Do they keep consuming the same amount and just pay more? Or do they actually reduce how much they’re consuming?

Knowledge at Wharton:  Are you able to glean enough from what has happened in places like Berkeley and Philadelphia to say, “Yes, absolutely, there’s no question that the economic and health benefits are there for people to stay away from sugary drinks?” And are they doing it?

Lockwood: Again, this is an insightful question that cuts to the heart of the issue. In many cases, we still need more evidence to know the optimal size of these taxes. There’s some initial evidence from the tax that was imposed in Mexico, and the one that was imposed in Berkeley a couple years ago, that does suggest people reduced consumption in response to these taxes. But the estimates of how much they reduced consumption are really wide.

Economists talk about that responsiveness in terms of elasticity. If you impose a 10% tax, then by what percentage do people reduce their consumption? Do they reduce their consumption by 1.5% or by 25%? It’s a huge range. If you take the middling estimates of those, and you think that it’s a 10% or so reduction, which is where a lot of economists at this stage think the value probably lies, then our initial estimates are that some positive tax, maybe even a little larger than the ones that have already been imposed, is probably optimal. But again, we’ll know more going forward when we see the effects of these bigger cities in the next few years.

Knowledge at Wharton: The taxes that have been put in place haven’t gone overboard, and they haven’t underdone it either. They’ve gotten it in the ballpark so that if there is any increase down the road, it may be a half-cent per ounce or so. The cities have done a pretty good job, for the most part.

Lockwood: I don’t want to go on record saying, “It should be exactly 3.25 cents per ounce,” because we’re still waiting to see the evidence come in. But [based on] the economic research so far, my guess is that somewhere in the range of 3 cents to 4 cents per ounce — rather than the 2 cents per ounce that we’re seeing now — would be in the ballpark. I would say cities, luckily, have been in the range of reasonable taxes so far, given the hazy nature of the estimates we have.

Knowledge at Wharton: In Philadelphia, it’s been well-noted that the mayor would like to use the money from the tax to help improve pre-K education. A big focus of your research is on the good that is potentially done by these taxes that goes back to the community.

Lockwood: One of the key questions is whether it’s helpful to offset the regressivity costs of these kinds of taxes by targeting the revenues back toward poorer communities or poorer consumers. Whether that’s effective or not ends up being a pretty technical theoretical question that we get into a fair amount in the paper. The upshot is that if the reason that poorer consumers are drinking more soda than richer consumers is just a difference in preferences, then it’s not all that helpful to try to target those benefits back toward poorer consumers. There might be other arguments for why it’s really beneficial to support pre-K education. But if that’s the case, then we should be doing it through income taxes or whatever, regardless. We shouldn’t necessarily tie it to the existence of a soda tax.

Knowledge at Wharton: The impact of higher prices and taxes on cigarettes to mitigate cancer and other diseases has been discussed for years. The fact that cigarettes have seen higher costs has slowed down some people, but it hasn’t stopped [people from purchasing them]. Part of this has to do with the attraction that people have to consuming these kind of products, correct?

Lockwood: I think that’s right. One of the other questions about soda is the extent to which soda consumption is addictive in the same way that cigarette consumption is. We know from other research that people are much more responsive to cigarette taxes, to the decision to purchase cigarettes, before they start smoking or right when they’re starting, rather than after they’ve been smoking for many years. Then, if the price goes up, they’ll generally just pay the higher cost for it.

For soda, the question is whether people who are high soda consumers continue to consume soda at the same rates, like cigarette smokers — or whether they actually switch to other beverages or reduce consumption overall. Even if they’ve been consuming soda for a long time.

Knowledge at Wharton: Not that it’s directly a part of your research, but part of the problem is the companies that are bringing these products forward and the impact they have through the marketing. To a degree, it’s an uphill battle for some of these people if they want to try to step away from soda. If you want to step away from cigarettes, it’s a much harder prospect.

“If the goal of a tax is to discourage consumption of something that’s unhealthy, then people will only reduce their consumption if they actually see that tax and feel it.”

Lockwood: One of the other policy ideas that we look at a little bit in this paper but don’t delve into too much — and I think there’s more interesting work to be done here — is other kinds of non-tax policies, like some of the advertising bans or pictures of blackened lungs on cigarette packages, for instance. These kinds of policies aren’t exactly taxes but are intended to reduce consumption of cigarettes. There’s been some discussion of similar kinds of policies for soda.

Those kinds of policies, too, can make more sense when you have these goods that are consumed more by poorer consumers. They can help dissuade people from consuming this stuff without actually taking money out of their pockets, right? The idea is that less marketing would cause poorer consumers to drink less soda, without necessarily having to pay more for the soda they are drinking.

Knowledge at Wharton: Is there bias in this process in general? You’re talking about more people of lower incomes being affected by sodas and cigarette tax, compared with people of higher incomes who have the ability to put in energy-efficient windows and refrigerators.

Lockwood: I think that’s right. Just like you mentioned with the kinds of subsidies for energy-efficient appliances, you can reframe everything we’ve said about sodas in that context and think about the unintended regressivity costs of those kinds of policies. Often, the benefits are accruing precisely to folks in the population who have higher incomes, who aren’t necessarily the ones you’re trying to help out the most.

You can do a similar kind of exercise where you say, “Well, this doesn’t necessarily have the redistributive benefits that we would otherwise hope for from a tax.” But it does have this corrective effect of getting people to consume more energy-efficient stuff, just like the soda taxes discourage people from consuming unhealthy stuff. And that corrective benefit has to be weighed against the regressivity cost.

Knowledge at Wharton: Going back to the soda issue, what really has been the impact from the tax? After it was implemented here in Philadelphia, there was a line item for it on your receipt so that it wasn’t just baked into the cost of a 20-ounce soda. That was an important piece to this.

Lockwood: Again, it’s too early to have much formal economic analysis out of this. We’ll see going forward what the overall effects on consumption and soda purchases actually are. But one of the interesting questions going into this was to what extent the soda tax — which is actually imposed on the distributors who supply sodas to grocery stores — would be passed on to consumers in the form of higher soda prices.

One of the interesting things we’ve seen so far is that it does look like stores are really trying to make a salient connection and say, “This is how much the cost of your soda has gone up because of this tax.” That’s pretty interesting. On the one hand, sometimes people are dismayed to see taxes being passed through to consumers. They’d like to see those taxes being borne by the firm or corporations. But on the other hand, if the goal of a tax is to discourage consumption of something that’s unhealthy, then people will only reduce their consumption if they actually see that tax and feel it. In this case, having that tax passed through to the consumer and being front-and-center on the receipt might be consistent with the apparent goals of the policy.

Knowledge at Wharton: With cigarettes, you’re not seeing that itemized receipt that says what you’re paying because of the tax. It could be a very important deterrent in this process.

Lockwood: It certainly could be. It’ll be interesting to see whether those line items persist going forward or whether that was a temporary move by the grocers to try to explain why we had this one-time cost increase. But if they stay there, and the tax ends up being fully passed on to consumers on into the future, I think it will be interesting to see if that actually helps with these benefits. Maybe this is the kind of thing that we should have been doing with cigarettes all along, really emphasizing how much the cost is increasing because of these taxes.

Knowledge at Wharton: How much interest in this research topic is there from the medical community?

Lockwood: I think the medical research community has a big part to play in this. Although this is mostly a theory paper, it identifies the key parameters or estimates that will govern what that optimal tax is. A lot of that research can beneficially come from the medical community. Things like, how much more does medical care cost when people consume more sugar versus less? How much do people seem to be taking those costs into account when they are making their consumption decisions?

Those are exactly the kinds of things that people are studying right now. And as I said, we have some initial estimates. But I think this is an exciting time both for economists and medical researchers because we will have better estimates of this shortly.

“What you’d like to do is to have people switch away from these sugary things toward other things.”

Knowledge at Wharton: We’ve seen growth in an area called behavioral economics. We want to understand the economic impact of our behaviors right now, whether they are positive or negative.

Lockwood: Exactly. This is a very exciting and vibrant area of economic research, where we relax the conventional economic model in which people are fully rational and fully take everything into account when they’re making every single decision they make all day. This area of behavioral economics allows for what many of us feel: That there are a lot of things going on. A lot of things are confusing, and you’re not always paying perfect attention to everything, including what the eventual health costs are of everything that you might engage in. When that’s the case, there are beneficial things that the government can do to help guide behavior or explain those costs.

Knowledge at Wharton: Starting to put these theories together, what potentially is the impact on the lower incomes in Philadelphia and Chicago and San Francisco, compared with where we will be going in the next five to 10 years?

Lockwood: I think a lot of it comes back to these empirical estimates of trying to see what the effects of these taxes actually are. In Philadelphia, if it turns out that people basically don’t change their soda consumption because they instead just drive across the Ben Franklin Bridge and buy their soda in New Jersey, then you don’t actually get any of the health benefits that I was just talking about. People still consume the same amount. They get diabetes at the same rate. The only thing that happens is, they waste more gas driving across the bridge. That would be a downside of this kind of policy.

Now, there are things you could potentially do to correct that, like cooperating with nearby localities and jointly setting taxes so there aren’t these big differences across local borders. But if it turns out that people just buy at their local store, and this has a big impact, then we’ll find out that this is a useful policy to be implemented at the city level.

Another question along these lines is how much people substitute other kinds of drinks or how their consumption behavior changes. One way in which Philly’s tax was distinctive is that it also extended to cover diet soda beverages. From the economic health perspective, that’s not obviously a great move in terms of policy design. Maybe there are some unintended health consequences of diet soda, too. But the estimates now suggest that those are miniscule relative to the negative consequences of sugar consumption.

What you’d like to do is to have people switch away from these sugary things toward other things — maybe toward diet soda, if that’s really what they want. Having a better sense of whether people just keep consuming their sugary soda because diet also went up [in price], or whether they instead switched to bottled water or something, will have an impact on whether other cities then think about imposing taxes across the board on diet beverages, too.

Similarly, there will be a benefit for other cities in understanding how to make the case for these policies to their constituents. One of the interesting things that we saw with the Philadelphia mayor’s proposal is it was tightly tied to these spending programs on pre-K in a way that some of the previous Philadelphia soda tax proposals were not, the ones that actually ended up failing. This is something that other cities I think have taken note of — that sometimes in making the case for these kinds of things, it’s really helpful to show how those funds would be spent and how they can be retargeted to the affected classes.