By Dyuti Khulbe/Public Praxis Forum

We tend to think of social problems as things that can be fixed once we understand them well enough. Poverty, crime, inequality, congestion, environmental degradation, poor public services, each can be studied, measured and turned into a policy problem. From there, it seems reasonable to assume that a sufficiently well-designed intervention should move things in the right direction.

Social life, however, rarely follows the logic of a carefully designed plan.

A policy enters a world that already has its own institutions, incentives, relationships and informal ways of doing things. The people affected by it may interpret it differently from those who designed it, while organisations responsible for implementing it have their own constraints and priorities. Even when everyone involved is acting in good faith, the eventual outcome can look quite different from what was originally intended.

This is the problem of unintended consequences. It is also one of the reasons why trying to change society is considerably harder than designing a policy on paper.

An Old Problem in Sociology

The idea has a long history, but sociologist Robert K. Merton gave it a particularly influential formulation in his 1936 essay on the “unanticipated consequences of purposive social action.” His concern was not simply that people make mistakes. Rather, deliberate actions can produce effects that were not part of the original intention, even when the action itself was purposeful and rational.

That distinction matters because it moves us away from an easy explanation of policy failure.

A government may introduce a regulation to address a particular social problem, and the regulation may have a clear objective and be based on substantial evidence. Yet once people begin responding to the new rules, new patterns of behaviour can emerge. Organisations may change how they operate, individuals may adapt to new incentives, and some effects may become stronger while others are weakened.

The resulting outcome is not necessarily the product of one mistake. It can emerge from the interaction between the intervention and the society into which it was introduced.

When the Measure Becomes the Goal

Merton’s idea becomes easier to see when we move away from abstract social theory and look at something rather ordinary: a performance target.

Public institutions need ways to assess whether they are achieving their goals. If patients are waiting too long for hospital treatment, for example, it seems perfectly reasonable to set a target for how long people should have to wait and then use that target to see whether things are improving.

England’s National Health Service provides a useful example. From the early 2000s, the government introduced increasingly strong targets around hospital waiting times, backed by monitoring and managerial pressure. Waiting times fell substantially, and research suggests that the targets played a role in producing those improvements.

But once the target became important for how hospitals were evaluated, something else happened. The waiting-time number stopped being simply a way of measuring performance; it became something hospitals had to achieve. That created incentives to organise appointments, patient flows and administrative practices around the target itself.

This could produce behaviour that improved the numbers without necessarily improving every aspect of the underlying problem. A hospital might prioritise patients in ways that helped it meet the required threshold, or change how cases were managed and recorded so that fewer patients appeared to be waiting beyond the target. Such responses could occur alongside genuine improvements in waiting times.

This is the problem captured by Goodhart’s law: when a measure becomes a target, it can stop being a reliable measure of the thing it was originally intended to represent.

The important point is not that hospitals or staff suddenly became dishonest. They were responding to a changed set of incentives. Once a particular number affected how an organisation was judged, that number naturally became something people paid attention to and acted around.

And this is where Merton’s argument becomes useful. The intervention changes the conditions in which people act, and people respond to those changed conditions. Their responses can then produce consequences that were not part of the original intention.

The same pattern can appear in many settings. Schools may begin concentrating on the indicators used to assess them; universities can reorganise activities around rankings and measurable outputs; businesses may prioritise quarterly targets over longer-term goals. In each case, the measure can gradually acquire a life of its own.

The broader lesson is that an intervention does not operate on passive subjects. People interpret incentives, adapt to them and sometimes find new ways of responding to them. Once that happens, the behaviour produced by the intervention becomes part of the story of what the intervention ultimately does.

When Individually Sensible Decisions Create a Collective Problem

The next step is to take that idea beyond organisations and look at what happens when many individuals respond to the same change at once.

Imagine that a government increases the price of driving into a busy city centre in an attempt to reduce congestion. For an individual driver, avoiding the charge might be a perfectly sensible decision: they could take another route, travel at a different time, or use another form of transport.

Now imagine thousands of people making similar calculations.

The combined effect may be quite different from what any one person intended. If enough drivers choose the same alternative route, that road can become congested. If large numbers change when they travel, traffic patterns may shift to different parts of the day. A measure designed to reduce one problem can therefore create new pressures elsewhere.

What makes this sociologically interesting is that nobody necessarily has to make a wrong decision for an unexpected collective outcome to emerge. Each person is responding to the circumstances in front of them, yet the combined result of those responses can produce a pattern that no individual intended.

This takes us beyond the idea that unintended consequences are simply the result of a badly designed policy. Sometimes the difficulty lies in the interaction between many individually reasonable choices. An intervention changes the incentives or constraints facing people, people adjust their behaviour accordingly, and those adjustments interact with one another.

That is one reason social outcomes can be difficult to predict. The result of collective action is not always contained in the intention behind the action, because people do not simply receive interventions; they respond to them.

One Problem Can Move Rather Than Disappear

There is another form of unintended consequence that appears frequently in public policy: the displacement of a problem.

An intervention can reduce a particular problem in one place, group or institution while shifting some of its costs elsewhere.

Environmental regulation provides many examples of this tension. A government can successfully reduce pollution from one source while economic activity, production or consumption patterns shift in ways that create different environmental pressures elsewhere. Housing policies can improve an area while also changing land values and creating pressures for displacement. Measures designed to make a service more efficient can reduce costs for one part of a system while creating additional burdens for another.

The point is not that every intervention simply creates another problem. It is that social problems are rarely isolated from one another.

Trying to change one part of a system can alter the conditions surrounding another part. The question, then, is not only whether an intervention achieved its immediate objective, but what else changed because it did.

This is also why unintended consequences are not necessarily negative. An intervention can generate unexpected benefits as well as unexpected harms. A policy may produce an outcome that nobody planned but that turns out to be valuable. The category itself tells us only that the outcome was outside the original intention; it does not tell us whether that outcome was good or bad.

But Was It Really Unintended?

This may be the more uncomfortable question.

We often describe an outcome as an “unintended consequence” after the fact, as though that settles the issue. But unintended does not necessarily mean unforeseeable.

A government may not want a particular outcome while knowing that it is a possibility. Researchers may have raised concerns before a policy was introduced. Communities may have pointed to risks that were not reflected in official assessments. Officials may have recognised a trade-off and decided that the expected benefits justified accepting it.

If the outcome then occurs, calling it an unintended consequence can be technically accurate while still leaving something important out of the story.

The more difficult questions are about knowledge and responsibility.

What was known? What could reasonably have been known? Whose warnings were taken seriously? Which risks were considered acceptable? And who ultimately had the authority to make that decision?

This distinction matters because otherwise the language of unintended consequences can become a convenient way of describing institutional failure without examining responsibility for it.

There is a difference between being surprised by an outcome and knowingly accepting the possibility of one.

That distinction is worth keeping in mind whenever we hear that a policy produced “unintended consequences.”

Does This Mean We Should Stop Trying to Change Society?

No. That would be a much too easy conclusion.

Collective action is necessary precisely because many social problems cannot be addressed by individuals acting alone. Governments regulate markets, build infrastructure, provide public services, protect rights and respond to environmental risks because leaving everything to spontaneous social adjustment is not a particularly convincing alternative.

The challenge is not to stop intervening. It is to recognise that intervention itself becomes part of the social environment.

Once we see this, policymaking starts to look less like solving a fixed problem and more like navigating a moving system. Evidence, planning and expertise remain essential, but they cannot provide perfect foresight. Good policy also requires some understanding of how people and institutions might respond once an intervention changes the conditions in which they operate.

It requires something that can be harder to institutionalise: the willingness to notice when the world is producing a different result from the one anticipated, and to take that information seriously.

Perhaps that is the real value of thinking about unintended consequences. They remind us that good intentions are only the beginning of social action. Between what an institution wants to achieve and what eventually happens lies a world of incentives, responses, interactions and choices that cannot always be controlled.

The aim, then, should not be to design policies with the illusion that every consequence can be predicted. It should be to build institutions capable of asking better questions before acting, paying attention to effects beyond the immediate target, and changing course when experience shows that reality is more complicated than the original plan.

Because trying to change society is never simply a matter of deciding what should happen.

It is also about understanding what people and institutions are likely to do once we change the conditions in which they act.

Featured image: Photo by Lukas Kienzler / Unsplash