In the collection Complex systems
How do traffic jams form with no accident or bottleneck?
Above a critical vehicle density, small fluctuations in speed grow instead of fading, producing a jam that travels backward along the road with no obstacle needed.
AI-drafted from web search results, then checked by a second, independent AI that read 3 of its 4 cited pages in full and found the thread supported. No person has reviewed it.

The short answer
These slowdowns are often called phantom traffic jams. They arise when traffic is dense enough that free flow becomes unstable. Small variations in how drivers brake and accelerate then grow rather than die out, and eventually a cluster of slow or stopped vehicles forms and moves along the road. Researchers describe this as a collective effect of many interacting vehicles, not the result of one obstacle. [1] [2]
A chain reaction of braking
The US Federal Highway Administration defines a phantom jam as a slowdown with no crash, lane drop, merge or event behind it. In its description, one or more cars at the front brake, the car following too closely brakes in response, and so on, producing a ripple that can leave cars further back at or near a full stop. It adds that larger gaps between cars make a phantom jam less likely. [3]
The critical density
Physicists treat vehicles as particles that each adjust speed to avoid hitting the one ahead. Small fluctuations always exist. At low density they disappear and free flow continues. If the average density exceeds a critical value, the fluctuations grow steadily and break the free flow. The jam then travels along the road without decaying. Real highway data show a similar threshold: the relationship between density and flow splits sharply into free flow and congested flow at a critical density that is almost the same across highways. [1]
Continuum models such as the Payne-Whitham and Aw-Rascle equations describe two competing effects. Preventive driving creates a stabilizing pressure. Drivers slowing at higher density, combined with a delay in adjusting to new conditions, is destabilizing. Above a threshold density the destabilizing effect wins and small perturbations grow. [2]
Evidence from a circular track
In a 2008 experiment reported in the New Journal of Physics, vehicles were set moving at the same speed, evenly spaced, on a circular road, at a density prepared for the onset of instability. A tiny fluctuation grew until a jam cluster appeared and propagated backward like a solitary wave, at about the speed of jam clusters on a highway. The authors present this as the first experimental evidence that a jam can emerge without a bottleneck. A summary in The Conversation describes the same study as using 22 drivers. [1] [4]
Jamitons and the role of bottlenecks
MIT researchers found that in inviscid models of this type, the instabilities grow into stable traveling waves they named jamitons. A jamiton has a sharp jump in density on one side and a smooth decay on the other. Their analysis shows jamitons travel slower than the vehicles, so drivers meet a sudden density increase and must brake hard. Such waves vanish only through strong smoothing, meaning extremely cautious drivers, or a drop in density. [2]
The 2008 paper argues that bottlenecks are triggers, not the essential cause. A bottleneck seems to cause jams because it raises density above the critical value. [1]
Sources
- Traffic jams without bottlenecks—experimental evidence for the physical mechanism of the (iopscience.iop.org) Found by source search. Not checked by a person.
- MIT Mathematics | Traffic Modeling (math.mit.edu) Found by source search. Not checked by a person.
- Recurring Traffic Bottlenecks (Fourth Edition) - Appendix D. Definitions and Traffic Bottleneck Typologies - FHWA Office of Operations (ops.fhwa.dot.gov) Found by source search. Not checked by a person.
- Don’t change lanes – the maths of holiday traffic jams (theconversation.com) Found by source search. Not checked by a person.
Questions to follow next
- Why do phantom jams move backward against the direction of traffic?
- Can driving behavior or technology reduce phantom traffic jams?
- How do the Payne-Whitham and Aw-Rascle models differ?