#112 2003 · Municipality of Drachten (traffic engineer Hans Monderman) · Public infrastructure / traffic safety

Drachten deleted the traffic lights, curbs and signs at its worst junction and crashes fell nine-to-one

the problem

A busy junction kept producing accidents no amount of signage, signals or enforcement fixed

background

Laweiplein, Drachten's main junction in the Dutch province of Friesland, carried roughly 22,000 vehicles a day through a standard signal-controlled intersection, and it produced accidents at a rate typical of that kind of junction: about nine per year through the 1990s. The standard playbook for a dangerous junction was more control — brighter signals, clearer signage, stricter enforcement — and every town in the Netherlands had already run that playbook without eliminating the accident rate, because drivers who trust the infrastructure to manage risk stop managing it themselves.

Hans Monderman, a Frisian provincial traffic engineer who had spent two decades studying accident data, had come to the opposite conclusion from his profession's default: that formal traffic controls let drivers relax into autopilot, and that autopilot, not a lack of signage, was the accident's real cause. Removing the controls was politically almost impossible to propose to a town council whose job was to look responsible for road safety — the electorate's assumption is that fewer signs means more danger, not less.

what everyone would do

Add more control — brighter signals, clearer signage, stricter enforcement — the standard response to a dangerous junction. It had already been tried everywhere in the Netherlands without eliminating accidents, because formal traffic controls let drivers trust the infrastructure to manage risk instead of managing it themselves, so adding more of the same category of fix only deepens the inattentive autopilot that was causing the crashes.

what they saw

Monderman saw that signals and signage weren't reducing danger, they were removing drivers' sense of personal responsibility for it — the false confidence a controlled junction creates is what let drivers stop paying real attention. The fix wasn't a better instruction, it was removing the crutch entirely, forcing every road user back into active, alert negotiation with the actual risk in front of them.

the move

In 2003 Monderman redesigned Laweiplein as a 'shared space': a single paved square with no traffic lights, no curbs separating cars from cyclists and pedestrians, and no lane markings, built around a small roundabout and fountains instead of a signal box. Drivers, cyclists and pedestrians were left to negotiate the crossing by eye contact and yielded right-of-way instead of by obeying a light.

why it works

Formal controls tell drivers the infrastructure has the danger covered, so attention relaxes into rule-following rather than active watching. Removing every explicit control eliminates that false confidence: with no light to obey and no lane markings to follow, drivers, cyclists and pedestrians have no choice but to make eye contact and negotiate right-of-way in real time, which raises everyone's attentiveness and produces genuinely safer, if slower and more uncertain-feeling, interactions. The discomfort people reported wasn't a side effect to be minimized, it was the same mechanism producing the safety gain — uncertainty is what keeps people alert, and removing it would remove the alertness that was preventing crashes.

the payoff

Annual accidents at the junction fell from roughly nine per year before the redesign to about one per year afterward, even as traffic volume held around 22,000 vehicles a day and average crossing time dropped from about 50 seconds to 30. Cyclist and motorist self-reported feelings of safety slightly declined even as the actual accident count fell — the discomfort was the mechanism, not a side effect of it.

where it breaks

The design only works where negotiation between road users stays genuinely manageable at the traffic volume and speed involved — beyond some threshold, uncontrolled interaction among many vehicles breaks down into gridlock or real unpredictability rather than heightened alertness, which is why shared space doesn't scale to a high-speed highway interchange. It also depends on the public tolerating the discomfort long enough for the safety data to matter politically: several shared-space schemes were later reversed after public pressure over subjective unease even where accident numbers had genuinely improved, since voters respond to how safe something feels, not just what the data shows. And it requires everyone involved to actually be capable of the informal negotiation the design demands — poor visibility conditions or road users unable to rely on eye contact are put at a real disadvantage a signal-controlled junction wouldn't create.

what came after

Laweiplein became the reference case for the international 'shared space' movement; Monderman went on to design more than 100 such schemes before his death in 2008, and versions were later built in the UK, Germany and elsewhere. Some shared-space schemes, including later ones in the Netherlands, were reversed after public pressure over subjective unease despite the safety data — a reminder that the mechanism works on drivers even when it unsettles the people voting on whether to keep it.