Road Diets: How Narrower Roads Have, in Fact, Made Traffic Flow Faster
A familiar urban experience is spending an hour in traffic on a road that appears to have ample space: several lanes are available, yet one turning vehicle, a lane merge, or a badly timed junction can bring the entire stream to a crawl. Cities have traditionally responded by adding capacity. However, transport research suggests congestion can also arise from how existing road space is organised. This explainer examines the evidence behind “road diets”, where through lanes are reduced and the space is reassigned, and why some such interventions have improved traffic flow and safety.
When the Road Creates Friction
Take a typical arterial road familiar across Indian cities: two or three lanes moving in each direction, motorcycles filtering between cars, buses stopping along the kerb, vehicles entering from side roads and a car waiting to turn across the traffic. The road may be wide, but the interaction among these movements determines its usable capacity at that point. Once one vehicle stops or changes lanes, others brake, move around it and merge back, creating successive disturbances that can travel backwards through the traffic stream.
In its most common form, a four-lane undivided road is converted into three lanes: one through lane in each direction and a central two-way lane for vehicles turning left. The space released from through traffic can also be allocated to cycle lanes, parking, pedestrian refuge islands or public transport infrastructure. The Federal Highway Administration’s research describes the underlying mechanism as reducing lane changes, speed differentials, and conflicts between turning and through traffic.
A road may technically have two lanes available in one direction, but if one is repeatedly occupied by turning vehicles, stopped buses or merging traffic, the second lane does not provide its theoretical capacity throughout the corridor.
Separating movements can therefore reduce disruptions that propagate through the traffic stream, even while reducing the nominal number of lanes available for through traffic.
What Happens When a Lane Disappears?
South Florida Avenue in Lakeland, Florida, offers a particularly useful case because the city measured the effect on traffic rather than relying on perceptions of congestion. A five-lane section was converted into three lanes, with one through lane in each direction and a central turning lane. After more than 18 months of monitoring, average vehicle speeds fell from 33 mph to 30 mph. Still, morning rush-hour travel times changed very little: the northbound journey became one second longer and the southbound journey four seconds shorter. Afternoon travel times did rise, by 14 seconds in one direction and 72 seconds in the other.
A vehicle travelling at a higher average speed can still spend more time completing a trip if it repeatedly encounters queues, turning vehicles and merging traffic. A lower average speed can produce a similar journey time when vehicles move more consistently.
Seattle’s Nickerson Street provides another example of the trade-off. The four-lane road was converted to one through lane in each direction, with a central turning lane, bicycle lanes and parking occupying the remaining space. One year after the intervention, annual collisions had fallen from a five-year average of 33.6 to 26. The share of drivers travelling at least 10 mph above the 30-mph speed limit fell from 17 per cent to 1.4 per cent in one direction and from 38 per cent to 1.5 per cent in the other, while traffic volumes changed by only about one per cent.
A Federal Highway Administration empirical-Bayesian evaluation of 45 road-diet sites across California, Iowa and Washington found a 29 per cent reduction in total crashes after four-lane roads were converted to three-lane configurations. The estimated reduction was 47 per cent in Iowa and 19 per cent across California and Washington, indicating substantial variation across settings but a consistent effect direction.
A separate Rhode Island study published in 2022 provides further support using an empirical-Bayes before-and-after method designed to account for selection effects and regression to the mean. Across the treated sites, total crashes fell by 29 per cent, while fatal and injury crashes fell by 37 per cent.
A central turning lane removes vehicles waiting to turn from the path of through traffic; fewer through lanes reduce opportunities for high-speed overtaking and weaving; and pedestrians crossing the road have fewer moving lanes to negotiate.
Changing Trade-Offs with a Rise in Traffic Volume
The Federal Highway Administration treats 20,000 vehicles per day as a useful screening threshold rather than a hard cutoff: roads at or below that level may be suitable candidates, but corridors near or above the threshold require analysis of peak-hour volumes, turning movements, signal spacing and access points. FHWA also documents successful road diets at volumes above 20,000 vehicles per day, including examples reaching 23,000–24,000 vehicles per day.
A 2024 study of five high-volume road diets in Los Angeles provides a useful test of what happens beyond this conventional threshold. The researchers examined corridors carrying more than 23,000 vehicles per day and compared them with 16 similar multilane roads that had not undergone a road diet. Average traffic speeds were 6.7 per cent lower during peak periods and 7.9 per cent lower outside them, while travel times were about 16 seconds longer per mile. At the same time, collisions, injuries and deaths were lower by between 31.2 per cent and 100 per cent, depending on the outcome measured.
Reducing lanes on a high-volume road can add some delay, but the increase may be much smaller than a simple calculation based on the loss of one lane would suggest, while the safety gains can remain substantial.
What Happens When Cities Remove Road Capacity?
Seoul offers a more dramatic test of the same idea. In 2003, the city removed the elevated Cheonggyecheon highway running through its centre. It replaced it with a restored stream and public space, despite concerns that the displaced traffic would overwhelm surrounding roads. Instead, traffic redistributed across the wider network as some travellers changed routes, departure times or modes. The case became an important example of “disappearing traffic”: when road capacity is removed, not every vehicle previously using that road necessarily appears elsewhere as additional congestion. This is why congestion cannot be understood simply as a fixed number of vehicles competing for a fixed amount of road space; the road itself helps shape the demand placed upon it.
Road diets therefore offer a narrower but important lesson for transport planning.
Before adding capacity, cities need to understand whether congestion stems from insufficient road space or from how existing space is organised. Sometimes, changing that organisation can achieve more than adding another lane.