Why the safest route is the one nobody takes

Two roads connect the same two points. One is a motorway with guardrails, lighting and a speed limit. The other is a narrow lane with blind corners. The motorway carries twenty times the traffic and produces more accidents per mile than the lane, which nobody chose.

What everyone sees

The lane’s low accident rate is attributed to its qualities: it is slower, it forces attention, it is inherently safer. The contribution of its emptiness — which is itself caused by the motorway’s attractiveness — is not counted. If everybody moved to the lane it would lose the advantage it never designed.

What is actually happening

Peltzman described risk compensation: when an environment is made safer, users adjust their behaviour to consume the added safety, often as increased speed or reduced attention. Adams generalised this to a risk thermostat: individuals maintain a target level of risk, and improvements in external safety are offset by changes in behaviour. Wilde’s risk homeostasis theory makes the same prediction. The motorway’s safety features attract volume and speed, both of which raise the absolute risk. The lane’s lack of features deters volume and enforces caution, both of which lower it.

Why it stays hidden

The paradox hides because safety is attributed to the road rather than to the interaction between the road and its users. An empty road with blind corners is not inherently safe; it is currently safe because its conditions deter the traffic that would make it dangerous. The feedback loop — danger deters use, low use reduces danger — is invisible to anyone who evaluates the road at a single point in time.