At some point in most policy conversations, we end up here.
Not because we’ve run out of ideas — but because one idea keeps feeling reassuring.
Punishment.
If a behaviour worries us enough, the thinking goes, then stronger penalties should put a stop to it. If the behaviour continues, maybe the punishment just isn’t strong enough yet.
It’s a very human instinct.
And it’s worth pausing to ask — not whether it feels right — but why it so often doesn’t seem to work the way we expect.
The appeal of punishment
Punishment has a certain clarity to it.
It feels decisive.
It signals seriousness.
It reassures us that lines are being drawn and enforced.
The logic is simple enough to feel obvious:
If a behaviour is bad enough,
make the penalty severe enough,
and people will stop doing it.
And sometimes, in very specific situations, that logic holds.
But when we look at punishment as a general policy tool — especially for persistent behaviours — things get a lot messier.
What punishment is actually good at
It’s worth being fair about this.
Punishment can work well when:
- the behaviour is occasional rather than persistent
- people have broadly similar levels of choice
- enforcement is consistent and visible
- and the cost of compliance is low
In those situations, penalties can genuinely influence behaviour.
The trouble is that many of the behaviours we end up punishing in public policy don’t fit that pattern.
They tend to be:
- ongoing rather than one-off
- shaped by circumstance as much as choice
- unevenly distributed across society
- closely tied to stress, opportunity, or survival
And that’s where the cracks start to show.
What usually happens instead
When punishment is used to try to eliminate persistent behaviour, a few things tend to happen.
Not all at once.
Not everywhere.
But often enough that the pattern becomes familiar.
Some people stop.
Some people don’t.
Some people adapt.
And it’s adaptation — not defiance — that reshapes the system.
Why behaviour adapts
As penalties increase, behaviour doesn’t disappear.
It adapts.
It becomes:
- more hidden
- more selective
- more dependent on intermediaries
- riskier for those with the fewest alternatives
People with resources find ways around the rules.
People without resources absorb the risk.
This isn’t about better or worse morals.
It’s about how punishment interacts with inequality.
Enforcement isn’t neutral
There’s another part of this that’s easy to overlook.
Punishment doesn’t enforce itself.
It relies on:
- detection
- discretion
- judgement
- and capacity
Someone decides who gets stopped.
Who gets searched.
Who gets charged.
Who gets warned.
Over time, punishment-based systems tend to concentrate their effects on the same groups — not because anyone planned it that way, but because enforcement follows patterns of visibility and vulnerability.
There are incentives at work here too, even if we rarely talk about them.
Police, regulators, and agencies are under pressure to show activity, compliance, and results. That naturally pushes enforcement towards cases that are easier to detect, easier to process, and more likely to stick.
Which means attention drifts towards people who are:
- more visible
- less able to contest decisions
- and less protected by time, money, or influence
None of this requires bad intent. It’s simply what happens when systems reward throughput and certainty rather than fairness or long-term outcomes.
The incentives don’t eliminate bias — they quietly systemise it.
That’s not a bug in the system.
It’s how the system operates.
The cost that we rarely discuss.
This is where the consequences start to pile up.
Punishment-based policy doesn’t just create rules — it creates entire support systems to carry those rules out:
Policing.
Courts.
Legal aid.
Prisons.
Probation.
Monitoring.
Appeals.
And then the knock-on effects:
- disrupted families
- reduced employment prospects
- poorer health outcomes
- higher chances of reoffending
None of these costs appear in the original policy announcement.
They arrive later.
In different budgets.
Usually described as pressure, not consequence.
The escalating punishment cycle
There’s a loop here that’s surprisingly easy to fall into.
When punishment doesn’t reduce the behaviour it targets, that failure often gets interpreted as evidence that:
- enforcement isn’t tough enough
- penalties aren’t severe enough
- or tolerance has gone too far
So the response is… more punishment.
At that point, the policy stops being about outcomes and starts being about resolve.
And resolve is a very difficult thing to measure — but a very expensive one to maintain.
Should we be asking a different question?
If we pause for a moment, a different question starts to feel more useful.
Not:
“How do we stop this behaviour?”
But:
“What actually happens when we try?”
Who adapts?
Who pays?
Who gets pushed into greater risk?
And which public services quietly pick up the bill?
Once we ask that, punishment starts to look less like a solution and more like a very costly way of rearranging harm.
So… are we saying we shouldn’t punish?
No — and this is where it’s worth slowing down.
This isn’t an argument against boundaries, accountability, or consequences. Some behaviours cause direct harm and need to be stopped.
What we’re questioning is something more specific.
Whether punishment can reliably do a job it was never designed for — especially when we use it as the primary tool for managing persistent human behaviour.
Because when punishment is asked to do that job, the outcomes tend to look like this:
- higher costs
- deeper inequality
- growing complexity
- and very little reduction in the behaviour itself
And that’s before we even get into real-world examples.
Which is exactly where we’re heading next.
Next, we’ll take this discussion and apply it to something concrete and familiar:
drugs — and what happens when we try to eliminate demand by criminalising it.
That’s where all of these patterns stop being theoretical and start becoming unavoidable.
(Onwards.)
Discover more from Hysnaps Politics, Gaming, Music and Mental Health
Subscribe to get the latest posts sent to your email.
