Journal · No. 004

The day Sweden switched sides

Christoph Svoboda · November 13, 2026 · 3 min

At ten to five on a Sunday morning in 1967, the few drivers allowed on Swedish roads stopped where they were and moved their cars to the other side. It was the most disruptive morning the country had ever planned for itself, and it turned out to be one of the safest.

Sweden had prepared for four years. Around 350,000 signs were replaced. Every intersection got a second set of traffic lights, wrapped in black plastic until crews tore the plastic off that morning. The public had voted against the change in a referendum, 83 per cent for keeping the left, and the government went ahead anyway.

The day produced 157 minor accidents, 32 of them with injuries. The Monday after produced 125, against a normal range of 130 to 198, and none of them fatal. Motor insurance claims fell by 40 per cent.

Then they came back. Within six weeks the claims were normal again. By 1969 the accident rate was where it had been before.

The signs did not wear off in six weeks. The lane markings did not fade. What wore off was the one thing nobody had installed: for a few weeks every driver in Sweden assumed his own reflex might be the wrong one, and knew that everyone else was assuming the same thing about themselves. Researchers later put it plainly. Perceived risk ran ahead of actual risk, and people drove accordingly.

Sweden did not get safer by changing the rule. It got safer by spending six weeks in a state where nobody trusted their own reading and everybody checked.

Most change programmes budget for the signs. Nobody budgets for the six weeks.

Most change programmes budget for the signs. Nobody budgets for the six weeks, and nobody notices when they end.

In the glossary: Shared ReferenceNovelty Budget