When should a government regulate its citizens’ behavior, and when should it use behavioral tactics to nudge them toward doing the right thing? Behavioral scientist Pelle Guldborg Hansen, of Roskilde University in Denmark, has spent his career integrating nudges into the wider policy universe. In “Making Healthy Choices Easier: Regulation versus Nudging,” a 2016 review in Annual Review of Public Health, Hansen and his coauthors argue that it’s not an either-or question. We asked him to explain.

This conversation has been edited for length and clarity.

When governments want to change their citizens’ behavior, they use something called an “intervention ladder” or “intervention staircase.” What’s that?

It’s a descriptive model of what goes on in the public sector as it faces a behavioral problem. The reason why this is called an intervention ladder is they’re moving through this standard process as they try to change the behavior of citizens.

The first option is that you do nothing. If that's not acceptable, then you would usually begin with providing information. This is where you see public campaigns addressed at enlightening the citizen about health dangers relative to the product, like “smoking kills, you will get cancer. ” It’s basically about informing the citizen about a health risk. But often people won’t change their behavior solely due to the provision of information.

So then you go to the next level, which is what I call persuasion — you try to convince the recipient of the information about why they should care. My standard lecturing example is a bit foolish: Somebody tells me that smoking kills, and I say, “Well, you know what, you have to die of something.” Then that person would say, “Yeah, but have you considered that when you’re dying, you’re dying in a very painful way. You will be at the hospital coughing blood on your kids, who are standing next to your bed crying, saying you will never see your grandkids.” This is the main effort of persuasive campaigns. You’re trying to create some sort of cognitive dissonance, to provoke the recipient.

And if that doesn’t work?

If that doesn’t work, you start to look at incentives. You’d say now they have got the information — and they also have formed the intention — but they’re still not doing the right thing. Usually you think first about providing a positive incentive like providing a rebate, or making smoking-cessation programs free. One might even pay them for quitting cigarettes. Then if that doesn’t work, you have the negative incentivization, where you have fines, for instance, or taxes — fat taxes, sugar taxes and so forth.

If none of that achieves your goal, what's the final option?

If that doesn’t work, then you hit them with legislation. Usually that’s expressed in terms of a ban of some sort — you’re not allowed to sell cigarettes to this age group. Or they fine you for driving too fast. But when you get that fine, it’s not just a tax. It’s something more. You’re not paying for being allowed to drive so fast, you’re paying because you did something wrong, so there’s this symbolic intervention saying, “We as a group or a society don’t like what you did, so we’re going to impose a punishment on you.” It has a symbolic nature.

The further you get up the staircase, the more invasive citizens will feel that you are. Very few people have anything against the state informing them, for instance, about climate change or certain health risks. They might disagree, but they’re not joining a public rally because someone gave information. As you move up the staircase, when you get to the persuasion campaign they can become offended. When you start to put in incentives, they start to become angry — not with the positive incentives, but with the negative ones. When you put a tax on cigarettes, beer, sugar, then someone will get angry. Then when you get to the bans, you really enter the red zone. People become more and more angry the further you get up this staircase.

Where do behavioral insights — that is, nudging — fit on this staircase?

The traditional intervention ladder works for problems where people behave rationally. When the air-raid sirens go off, or the government informs about a gas leak in our neighborhood, information usually does the job. But there are also behavioral problems, where these standard approaches don’t work very well. A behavioral problem is a tendency or pattern of behavior where the agent has information, attitudes and incentives to do otherwise but still doesn't do it. For instance, smoking. I know that it kills, I might even have an attitude that I want to stop, there’s a huge tax on cigarettes — but I still don’t quit smoking. There’s basically a behavioral problem. And that’s what we study with behavioral insights.

“We have biases in decision-making, biases in intention, and biases in self-control.”

Pelle Guldborg Hansen

When you work with behavioral insights, you step away from the staircase of traditional regulation and you say, “Let’s try to look at the problem as a different one. Let’s try to look at this behavior not as a rational problem of information, attitudes and incentives, but as a problem of irrationality — that is, a behavioral problem.” We have biases in decision-making, biases in intention and biases in self-control. When you take these standard regulatory tools of information, persuasion, incentivization and legal approaches, and apply them to behavioral problems, you end up with some inelegant solutions, like an extreme tax on cigarettes, which could have been handled differently if you used behavioral insights. E-cigarettes would be a good candidate. When it comes to habits, it's much easier to implement a substitute than to kick the habit. If e-cigarettes are less harmful, encouraging current smokers to shift to them might be a much more effective strategy. [Editor’s note: Here’s more discussion about e-cigarettes, both pro and con.]

So nudges aren't one rung of the intervention ladder?

You usually don’t place the behavioral interventions at a position on this intervention ladder. You apply them at any rung of the ladder so that each intervention works better. Suppose the information campaign and the persuasional campaign aren’t working, and now we’re considering moving on to incentives. But perhaps instead, we could tweak the way the information is presented, or change the persuasion strategy to have a better result. That way, we don’t have to move further up the intervention ladder to more extreme measures. That way, we avoid pissing people off.

“If you combine the traditional regulatory measures with behavioral insights, you get a stronger policy.”

Pelle Guldborg Hansen

So in your way of thinking, it’s not an either-or situation, where you choose between nudging and regulation. They’re really two independent approaches, aren’t they?

What I’m saying is we need to combine. If you’re going to introduce some incentives, then you can combine this with nudges, with behavioral insights, to boost the effect. If you’re going to provide people with information, then if you frame the message with behavioral insights in mind, information might be all it takes, and you don’t need incentives. The point is, if you combine the traditional regulatory measures with behavioral insights, you get a stronger policy.

If people don’t pay their taxes, they will be fined at some point. But if you combine that with behavioral insights about social norms — sending a letter to remind people that most of their neighbors pay their taxes on time, say — you get more people filing their taxes on time. It isn’t a choice between one and the other.

Critics sometimes object that nudging can be used as an excuse for deregulation, for having the state back out of its responsibilities. Is that a valid criticism?

Nudging is not an excuse for rolling back the state. It’s an argument for influencing behavior in a smarter way based on what we know about human behavior. If we can create a behavior change just by relying on a tweak of information or persuasion, or a tweak of the way we design incentives, then why shouldn’t we do that? Then we don’t annoy people without any good reason.

On the other hand, other critics object that governments are overreaching by manipulating their citizens through nudging.

That’s not a criticism you can post against nudging as such. Working from a state perspective, we’re always trying to influence people. This is what the state does — it tries to influence behavior. That’s not about nudging, that’s about regulation. Using behavioral insights is just another kind of regulation.

Sometimes critics ask, “How do you know what people really want to do? If you can’t do that in any reliable way, doesn’t nudging have a problem?” The answer is no, it’s not nudging that has a problem — it’s all kinds of policymaking. It’s a political question of what we should try to do, which goals we should try to achieve. And then, if you have good politicians, they ask their ministries: “How do we do this in a gentle, cost-effective way?” The choice of which goals we pursue with nudging isn’t relevant to evaluating nudging as a tool.