Why Effectiveness and Ethics in Behavioral Design Are Not Friends

I am currently working on a project in the field of preventive health research. And precisely there, perhaps more than anywhere else, I encounter a question that the applied behavioral sciences consistently avoid: What happens when effectiveness and ethics in design cannot both be achieved at the same time?

Prevention is by definition an intervention in behavior before a problem exists that the affected person perceives as a problem themselves. This makes it perhaps the clearest test case for a conflict that is happily hidden in policy design, but is hard to hide in medical contexts: Who decides which behaviors are desirable? And whose conception of a good life underlies this judgment?

What follows is not a plea against behavioral design. It is an attempt to ask an uncomfortable question honestly.

The Workshop Question That No One Answers

There is a conversation that takes place in workshops on behavioral design, nudging, and organizational development with remarkable regularity. It usually begins with someone presenting a case study in which behavior was successfully changed through clever design, whether it's a higher savings rate through automatic enrollment in retirement programs, healthier food choices through changed cafeteria architecture, or higher vaccination willingness through changed appointment allocation. The initial reaction in the room is enthusiasm. Then inevitably comes the hand. And then comes the question, which is always the same, even if it takes different forms: But isn't that also a bit manipulative?

The answer typically given in these workshops is: No, because we leave the person with a choice. We only change the context. We nudge, we don't force. This is the famous libertarian paternalism of Thaler and Sunstein, the elegant formula that promises effectiveness and ethics simultaneously, without having to choose between them.

The Power Problem Disguised as Technology

Here begins the real problem, which goes far beyond the technical question of whether a particular nudge works or not.

Politics has this problem on a large scale. Whoever sets incentives, defines default options, constructs subsidy premiums, and shapes tax breaks inevitably makes a prior decision about which behavior counts as desirable, which lifestyle is deemed worthy of promotion, and which future is desirable enough to be steered toward with state resources. This is not a neutral technical operation. It is a value judgment disguised in the language of efficiency, because the language of value judgment is politically more awkward.

Take the discussion of climate policy behavior control. The state that offers purchase subsidies for electric vehicles, creates tax incentives for energy-efficient renovation, and levies charges on airfare claims to merely set frameworks. It claims to change only the economic context. The citizen makes the decision themselves in the end. This is formally correct. It is also a form of self-beautification. Because whoever defines the frameworks also defines which decisions are easy and which are hard, which lifestyles swim with the current and which must struggle against it. That is power. And power that disguises itself as mere context design is, in a sense, an even more interesting form of power, because it is more invisible than open regulation.

The Unavoidable Trade-off

Now one might object that this is unavoidable. That every political decision, even the decision to do nothing, sets a framework. That neutrality is an illusion and that the question therefore is not whether we intervene in behavior, but only how and for what purpose. That is correct. But it is precisely at this point that most discussions about behavioral design fall silent, because the consequence is uncomfortable.

The consequence is this: Effectiveness and ethics in behavioral design are not natural allies. They stand in a genuine, structural conflict. And this conflict cannot be resolved. It can only be consciously carried or unconsciously suppressed.

What does this mean in practice? It means that every effective behavioral design is by definition asymmetrical. It facilitates certain decisions and makes others harder. It favors certain motivation structures and neglects others. It is calibrated to a target audience that is implicitly set as the norm, and everyone who doesn't fit that norm experiences the design as friction, as resistance, as the nagging uncomfortable feeling that the system somehow wasn't built for them.

Whoever denies this, either because they don't want to see the complexity or because they can't afford to see it for strategic reasons, produces behavioral design that appears effective if you measure the right metrics, and simultaneously systematically renders certain groups, motivation structures, and life realities invisible.

Organ Donation as a Case Study in Conflict

An example that illustrates this tension so well that it sparked a genuine political debate in Germany: the question of organ donation. Opt-in or opt-out?

In an opt-in system, only those who have actively consented donate. In an opt-out system, everyone who has not actively objected donates. The behavioral economics finding is clear and has been replicated for decades: the default option determines the outcome to an extent that destroys any illusion of neutral decision architecture. Countries with opt-out regulations have dramatically higher donor rates. The design works.

But what actually works here? It works through inertia. It works through cognitive convenience. It works through the human tendency not to change the default state, not because you consciously prefer it, but because objection requires effort, effort requires attention, and attention is a scarce resource. The opt-out design exploits precisely this resource scarcity to produce behavior that is socially desired. It is effective. It demonstrably saves lives.

It is also an intervention in bodily autonomy through the mechanism of inertia, which enrolls people who don't actively object, not because they've consented but because they haven't organized the objection, in a system that exercises authority over their bodies after death. For people with stable health, sufficient time, and administrative competence, objection is a trifle. For people living in precarious conditions, who have no relationship to state form processes, or who have deep cultural or religious objections to organ donation but don't know how to object, the opt-out mechanism is something else.

The opt-out design is effective for the target audience the designer implicitly had in mind. It is paternalistic toward everyone else. Both at the same time. Without resolution.

Asking the Uncomfortable Question

The question that arises from this is not technical in nature. It is ethical in nature. And it does not ask: How do we design something that is both effective and completely ethically unproblematic? This question is poorly framed because it implies an answer that doesn't exist. The correct question is: Which side of this conflict do we consciously carry, and what consequences do we accept for it?

This is an unfamiliar formulation in a field that likes to speak in the language of optimization. Optimization implies that there is a solution that maximizes multiple goals simultaneously. The reality of behavioral design is that the most interesting and consequential decisions are not optimization problems, but trade-off problems, in which you must decide whose interests, whose motivation structure, whose conception of a good life you place at the center, and whose you inevitably give less attention to.

Why This Becomes Even More Invisible in Organizations

Politicians who can state this honestly are rare, because honesty about trade-offs is structurally disadvantaged in democratic communication systems. The message "we have decided that this conflict is resolved in this direction, and this has these consequences for these groups" is harder to sell than "we have developed a program that works for everyone". The second message is almost always false. The first is almost always politically risky. So the decision about the trade-off migrates into the technical language of design, where it becomes invisible and therefore no longer needs to be justified.

This may be the real democracy problem in behavioral design, not that nudges exist, not that the state sets frameworks, but that the value decisions embedded in these frameworks are systematically translated into a language that makes value decisions invisible. Psychology becomes explained as technique. Politics becomes design. And the question of whose behavior is being changed in whose interest stops being a legitimate public question because it has disappeared into the jargon of behavioral architecture.

Now the political level makes this problem visibly large, because the consequences are large and because there is at least the possibility of democratic control. There are elections. There are parliaments. There is, theoretically, accountability.

And then there are organizations.

In companies, in institutions, in the thousand everyday contexts where HR departments design incentive systems, leadership development practitioners build feedback architectures, and cultural programs define which behaviors count as "engaged," "high-performing," or "culturally fit," the same process occurs. Effective behavioral design that favors certain motivation structures, implicitly sets certain lifestyles as the norm, and hides the trade-off between effectiveness and autonomy in the language of culture, values, and development.

The difference from politics is not that the trade-off is smaller. The difference is that the language in which these decisions are made is even further removed from an honest naming of the conflict.

So whoever asks which side of the conflict between effectiveness and ethics they consciously carry and what consequences they accept for it is not asking only politicians and policy advisors. They are also asking themselves.