Traffic circles, roundabouts, rotaries

Last fall the Economist had a short feature on traffic circles that I only noticed today. I was about to riff on how wonderful they are, and how underused in most of the US, but I’m already walking back the post I had in mind.  For drivers, and people who breathe air, they are pretty much all win:

The Insurance Institute for Highway Safety, an independent research group, estimates that converting intersections with traffic lights to roundabouts reduces all crashes by 37% and crashes that involve an injury by 75%. At traffic lights the most common accidents are faster, right-angled collisions. These crashes are eliminated with roundabouts because vehicles travel more slowly and in the same direction. The most common accident is a sideswipe, generally no more than a cosmetic annoyance.

What locals like, though, is that it is on average far quicker to traverse a series of roundabouts than a similar number of stop lights. Indeed, one national study of ten intersections that could have been turned into roundabouts found that vehicle delays would have been reduced by 62-74% (nationally saving 325,000 hours of motorists’ time annually). Moreover, because fewer vehicles had to wait for traffic lights, 235,000 gallons of fuel could have been saved [MO: with attendant criteria pollution and noise reduction].

They have three defects, though, one tractable and two not so much.  The easy one is that they reverse the convention of priorité à droite, the obligation of the car at the left of a merge to yield the right of way to the car on the right (this post assumes right-hand driving). This convention is not a problem when a minor street enters an arterial or avenue at right angles from the right (especially if it has a stop sign or a light), but the whole idea of a circle is that the circle itself and its tributaries have equivalent status and acute-angle convergence.  In the new traffic circle, very rare in California, just set up in Berkeley,  a few months of use and some yield signs seem to have done the job and we are zipping around and through it, correctly yielding to the traffic already in the circle.

Good as they are for vehicle traffic, however, they are really bad for pedestrians.  Notice the Economist’s careless implicit definition of “locals” as “people who drive in the neighborhood”.  Walking around a traffic circle of any size is either a long detour, crossing tributaries whose whole spirit is that their flow never stops,  or navigating a one-way whirl, twice, that never has a red light and may be two or more lanes wide.  When tributaries are light-controlled, things are almost worse: how do you get across DuPont Circle on your feet?…and that one even has a tunnel under it to divert some of the vehicle traffic.  The aerial view in the Economist article rewards close attention: the sidewalks in this suburban wonder peter out in confusion and gravel and actually disappear as they get near it. Obvously the traffic engineer had no idea what to do next, and with good reason.

Circles larger than the little three-meter traffic calmers that substitute for (or, wrongly, complicate) four-way stops at minor intersections take up a lot of space, aggravating the effect of roads in driving origins and destinations further and further apart and causing sprawl.  This space can be landscaped, but not used: the larger the circle the more unrelenting and unforgiving the traffic going around it that separates it from pedestrian life.

Too bad: another really seductive idea with fatal loose ends and crippling baggage.

Author: Michael O'Hare

Professor of Public Policy at the Goldman School of Public Policy, University of California, Berkeley, Michael O'Hare was raised in New York City and trained at Harvard as an architect and structural engineer. Diverted from an honest career designing buildings by the offer of a job in which he could think about anything he wanted to and spend his time with very smart and curious young people, he fell among economists and such like, and continues to benefit from their generosity with on-the-job social science training. He has followed the process and principles of design into "nonphysical environments" such as production processes in organizations, regulation, and information management and published a variety of research in environmental policy, government policy towards the arts, and management, with special interests in energy, facility siting, information and perceptions in public choice and work environments, and policy design. His current research is focused on transportation biofuels and their effects on global land use, food security, and international trade; regulatory policy in the face of scientific uncertainty; and, after a three-decade hiatus, on NIMBY conflicts afflicting high speed rail right-of-way and nuclear waste disposal sites. He is also a regular writer on pedagogy, especially teaching in professional education, and co-edited the "Curriculum and Case Notes" section of the Journal of Policy Analysis and Management. Between faculty appointments at the MIT Department of Urban Studies and Planning and the John F. Kennedy School of Government at Harvard, he was director of policy analysis at the Massachusetts Executive Office of Environmental Affairs. He has had visiting appointments at Università Bocconi in Milan and the National University of Singapore and teaches regularly in the Goldman School's executive (mid-career) programs. At GSPP, O'Hare has taught a studio course in Program and Policy Design, Arts and Cultural Policy, Public Management, the pedagogy course for graduate student instructors, Quantitative Methods, Environmental Policy, and the introduction to public policy for its undergraduate minor, which he supervises. Generally, he considers himself the school's resident expert in any subject in which there is no such thing as real expertise (a recent project concerned the governance and design of California county fairs), but is secure in the distinction of being the only faculty member with a metal lathe in his basement and a 4×5 Ebony view camera. At the moment, he would rather be making something with his hands than writing this blurb.

45 thoughts on “Traffic circles, roundabouts, rotaries”

  1. Are there any reports of how the German town that abolished lights and if I remember, stop signs, has worked out now that a year or more has passed?

    1. I mean the circle at Piedmont and Channing, which just got its stop signs replaced with the correct yield instructions.

  2. The roundabout is essentially a system of priorities; it need not, and in Britain often is not, a physical object more complicated than a simple marking on the road surface. So it can be combined with a light-controlled X pedestrian crossing.
    Roundabouts also cause problems for cyclists.
    I’m puzzled by your Francophile priorité a droite convention. The British Highway Code does not, as I recall from age 17, even consider the possibility of an unmarked road junction. Driving along a road, you have priority until you hit a sign saying you don’t.

    1. Yielding to the car on the right is most important on multi-lane roads with lane switching and weaving, and many intersections in the country (and in the city, until they get stop signs or lights). When two cars arrive at an intersection approximately together, it’s useful to know that the one on the right should be allowed to go first.

          1. To see how Stalinist Britain handles this, look at the current UK Traffic Signs Manual. Sections 3:3.8, 3:3.13 and 5:3.15 provide for universal Stop or Give Way signs or painted road markings, the exceptions being “private accesses or minor estate [JW: housing estate] roads where traffic speeds and flows are low and visibility is good.”
            The level of detail is extraordinary. Road signs should be mounted at 95 degreea from the road except on bends, section 1:1.51, to reducet reflected glare from headlights.

        1. Well, technically, in Michigan you can drive from Timbuktu to Kalamazoo. 🙂

          More seriously, the Michigan Vehicle Code (like, I believe, all others in the United States) does indeed provide that you have to yield to the car on the right at intersections (assuming that other rules don’t apply). At least that’s what I learned when I was preparing for my driver’s license, and the law says so, too; it’s also mentioned in “What Every Driver Must Know”, the booklet that the Michigan Secretary of State provides to people taking a driving test.

          Do there, in practice, exist unmarked intersections in Michigan? I can’t actually recall a single one now that I think about it — outside of very, very rural areas that the Michigan Department of Transportation may have forgotten about — and I lived there for the majority of my adult life. I think the law is primarily there as a fallback regulation (as opposed to, say, Germany, where you practically can’t seem to drive through some suburbs without yielding at every intersection). [1]

          As to driving in the UK: What I was told when learning to drive here (aside from everybody seemingly speaking a totally different language, e.g. “priority” instead of “right of way”, “road junction” instead of “intersection”) and inquired about unmarked intersections was that nobody had priority and everybody should proceed with caution (same as at a US 4-way stop intersection).

          [1] I would still be careful about assuming that my right of way at an unmarked intersection is honored in Michigan; just as any pedestrian in Michigan should be very careful about naively assuming that drivers consider crosswalk markings to be anything but ornaments with no practical purpose.

          1. In my medium-sized city (Madison, Wisconsin), innumerable low volume intersections, mostly in residential areas, are unmarked. No prob.

            Traffic circles are relatively new hereabouts, but people get used to them reasonably quickly. My take, from experience and reading, is that when intelligently designed for well-chosen sites, they are an improvement from both safety and efficiency perspectives. We have competent traffic engineers on the public payroll who, I am confident, have given quality consideration to all points raised in this discussion

          2. Priority at a 4-way stop exists (in CA at least) – if two or more cars arrive at the same time, the car(s) to the right have priority.
            Can be a bit confusing, as it involves actual agreement about who is there first, so it’s more a rule of thumb than anything.

        2. In my neighborhood in a California city of not quite 100,000 residents, there’s a multiblock stretch with no stop signs at the intersections. It felt daring as heck the first few years, but now I’m used to it. Never seen a wreck yet, but my doctor’s office is in the neighborhood too, and he says he’s seen a few gnarly ones. Oddly, though, the city has never signed them.

  3. Circles are a problem for peds, but I think not as much as O’Hare thinks. The big problem with circles is getting used to them. For drivers, it means learning a new set of conventions. For peds, it means planning your route to avoid them. I have often traipsed around Dupont Circle, in DC. It’s seldom a problem, because I am usually trying to get from Point A to Point B, neither point quite on the circle itself. You can usually avoid circumnavigating the circle if you pre-plan.

  4. Hanover, NH installed 2 of these a few years ago when it built a new middle school on the main route north of town. They bracket the middle school, about half a mile apart. One of them is purely for the purpose of slowing traffic coming from the north (from a stated speed limit of 50 mph — actually 35 at that point, but before the circl considerably higher — to one of 30 mph, 20 mph when kids are being dropped off or picked up); that is, only 1 road goes in &/or out of the circle, route 10. The other is where the road from the grade school joins the main road, on the way out of town; route 10, plus another street that terminates at the cirle. This used to have a blinking yellow light. The northern circle has 2 exits/entrances, and the southern one has 3. In this part of the country, pedestrians have absolute priority on any cross-walk that has no light, and people are excellent about observing this (although the pedestrian does have to be clear about intending to cross — no hanging around on the curb is going to stop cars, you do have to step into the street) so crossing on foot is not a problem.

    I’m not clear on the priorité à droite bit; priority is first to pedestrians crossing, and 2nd to any car that is in the circle already.

    Some adjustment was necessary initially, but IIRC, it didn’t seem to take more then 2-4 months before people around here were clear on the concept.

    1. The Dutch interwoven light-controlled fully separated systems for feet, bicycles, trams and cars is a wonder to behold and a pleasure to use…though it would be a lot less land hungry without the fourth feature.

  5. I think it’s a mistake for the writer to conflate modern roundabouts with huge, City-Beautiful-era traffic circles like Dupont Circle. They are really two different creatures from a traffic engineering and roadway design perspective, and have very different functional attributes and advantages/disadvantages, as well.

  6. One traffic circle maneuver that always seems terrifying is exiting the circle from the inner of two lanes. I admit to not having seen this cause any collisions, but often this is only due to luck or great alertness on th epart of drivers in the outside lane.

    How, according to the system, is this problem supposed to be solved?

    1. try it on a motorcycle, if you want to really taste fear. One way to do this is to keep going around and around until late at night when there’s less traffic…

    2. A properly designed two-lane roundabout [1] should also have two lanes entering and exiting at each side.

      I’ll be assuming that you’re living in a country where you normally drive on the right.

      If you plan on leaving the roundabout at the next exit, you enter the roundabout in the right lane, stay in the outer lane, and exit taking the right lane again.

      Otherwise, you enter the roundabout in the left lane, stay in the inner lane until you exit, and exit taking the left lane.

      This way, you only have to yield right of way when entering a roundabout; you should never need to change lanes within a roundabout or cross paths with another vehicle.

      [1] “Properly” designed does not mean that there can’t be variations on this theme, especially depending on local traffic patterns.

      1. That works if all drivers understand that the right lane is reserved for those who plan to exit at the first opportunity after entering. My experience suggests that this rule is not widely obeyed, at least not around Boston. Even where it is followed, it implies that in a roundabout with, say, four entry/exit roads two-thirds of the drivers must travel in the left lane. Is that a good idea?

        1. You solve the first problem by road markings and signs that direct drivers accordingly.

          The perceived 2:1 imbalance may or may not be a problem. Having the same number of cars in the left and right lane is worse if cars constantly have to stop rather than guaranteeing an uninterrupted traffic flow. Having twice as many cars in the left lane is of no concern at all if, as a result, fewer cars have to actually wait.

          If the traffic pattern requires it, other approaches can also be similarly safe, if the roundabout is designed accordingly. For example, a common variation is that the right lane can go straight ahead or right, while the left lane can go straight ahead, left, or do a U-turn. In this scenario, cars that are in the right lane and go straight ahead need to yield to traffic in the inner circle leaving to the right (which can be encouraged through appropriate road markings). However, the so-called turbo roundabouts is probably the superior solution to the two-lane roundabout problem in that case.

  7. Where in Berkeley is this new traffic circle? I don’t recall hearing about it and I live in Oakland.

    1. The circle at Piedmont and Channing is, um, where Piedmont crosses Channing. It’s physically a longtime feature of the Olmsted mini-parkway design of Piedmont, but mis-signed with 4way stopsigns until a few months ago.

      1. Oh, ok. I know the place, but as you say it didn’t really have a traffic-circle feel. Just a regular 4-way intersection with this weird thing in the middle you had to drive around. I didn’t know they changed it.

      2. Those 4-way stop signs are not the only example of a poorly setup circle, but were a constant source of amusement and confusion to me the year I spent at Berkeley. Actually 4 way stop signs are generally confusing. Far better (in terms of vehicle wear and fuel use) to have a two way stop and allow one road to pass through unhindered.

  8. Traffic circles are nothing new to the U.S., being, as noted above, historic design elements of the District of Columbia’s orignial L’Enfant plan where they were meant to provide cannon emplacements with a clear shot from Dupont Circle up, say, Connecticut Ave. (beware, Marylanders!) in case of invasion or civic unrest.

    As someone who lived in the ‘hood while in grad school, I assure you that Dupont Circle was and is not a pedestrian barrier what with its lovely fountain, pathways, chess tables and flowers. Indeed, it is a pedestrian magnet and I would much rather walk through it than around it. It does benefit from a variety of traffic calming features such as concrete lane barriers in car part of the circle, lots of traffic lights and endless congestion, somewhat alleviated by the tunnel carrying Conn Ave traffic under the Circle. Indeed, lots of D.C.’s historic circles have had underpasses added due to the horrific congestion they cause, e.g., Washington Circle.

    And, they have been very popular in New England for many years where they are known as roundabouts or rotaries. See, e.g., and I personally have some fear about confronting Boston drivers at a rotary.

    And, having been trapped for quite a while trying to escape from the world’s best known rotary at Etoile where twelve avenues and every card in France come together I know that there are some problems–I personally have sworn to never again enter Etoile as a driver.

    Finally, I found that the endless roundabouts the driver encounters in the country in Southern France, Spain and Portugal to be irritating at best. The first two or three okay, but, when you hit twenty in an hour it can get to you, not to mention slowing you down–a lot. Of course, there is something to be said for slowing down French (and Bostonian) drivers.

    Pedantic footnote: L’Enfant’s plan for the District of Columbia is way before the City Beautiful movement.

    1. I believe that the Étoile roundabout in Paris has now had its rules changed so that incoming traffic has to yield. It used to be strictly priorité à droite, and was scary as hell for the uninitiated (and maybe for regulars too.) The place de la Concorde used to be another, not quite as unruly as the Étoile, but it now has traffic lights at strategic places. I rode a bicyle around it at rush hour last October, feeling quite safe. If you had told me before I saw the new set-up that I would ever contemplate such an adventure, I would have called you crazy.

      And driving in the south as part of the same stay in France, I was delighted to have ambient traffic slowed … as H says.

  9. Circles can be a problem for pedestrians, so the trick is to give peds protected alternatives nearby. The Belisha beacon crossings are one sort. In the states, there are marked pedestrian crossings, but they seem to be honored more in the breach than in the observance, at least in my midsouthern, moderately upscale neighborhood.

  10. If the roundabout is in a really inconvenient place for pedestrians, you can always put subways under it. I don’t know how much it costs to dig a pedestrian subway though – probably enough that it’s only worth doing for major junctions in very densely-populated areas. In London at least, some monuments and underground stations are located at the centre of roundabouts, which gets the most out of the subways.

    1. Taking pedestrians off the streets (pedestrian underpasses/subways) is exactly the wrong way to manage what should be a lively concentration of activity. Getting up and down in them is not nice for the wheelchair bound, or the old; with elevators they get quite expensive.

  11. in texas, many years ago, it was relatively common for smaller towns with an intersection of 3 or more larger roads to have a traffic circle because the city didn’t have the recurring cost of the electricity for a traffic light. as town revenues and federal subsidies to the states increased, these traffic circles were eliminated and replaced with signal lights. there are several towns i traveled through frequently with my parents that had these circles. i always thought they were exciting but my parents weren’t big fans of them.

  12. Michael: “The aerial view in the Economist article rewards close attention: the sidewalks in this suburban wonder peter out in confusion and gravel and actually disappear as they get near it. Obvously the traffic engineer had no idea what to do next, and with good reason.”

    In the USA at least, a lack of sidewalks does not indicate confusion on the part of traffic engineers. It indicates an attitude of ‘f*ck pedestrians’.

  13. I live in Queensland, Australia where we have had roundabouts for more than 20 years, and yes, they can be very useful in light to moderate traffic. even then it can be a problem trying to enter the roundabout from a minor road. It can take some time for a break in traffic from the major road sufficient to allow access. In heavy traffic (2 major roads) it can be quite stressful waiting for the break that will let you in. In many cases the traffic authorities have dramatically increased the size (diameter) of the circle such that it becomes simply a 2 way intersection, the entry points and exit points so far apart that they become another intersection. Even, in many instances, adding lights to each of the created intersection.,+Queensland&hl=en&ll=-26.66107,153.066663&spn=0.003078,0.006539&sll=-25.335448,135.745076&sspn=49.716039,107.138672&oq=mar&hnear=Maroochydore+Queensland&t=h&z=18
    Sorry I do not know how to create this as a link

  14. If you want to walk down Connecticut Avenue through Dupont Circle, you have to cross six separate roadways to do so. But Dupont Circle is one of the more pedestrian-friendly circles in Washington. Ward Circle may be the worst.

  15. I live in Bend, Oregon where we embraced roundabouts. I love em. However there are a few drawbacks: Land constraints sometimes limits the size(circumference)and results in limiting traffic flow. Be aware of pedestrians and bicyclist (bicyclists who don’t follow the rules of the road my particular pet peeve). Drivers have to be aware that the roundabouts create a different relationship with pedestrians and cyclists than the visual orientation of a normal intersections. Different window post blind spots. Tourists – Harrumph! Dealing with drivers who don’t understand the roundabouts can be frustrating, but be assured you only run into these divers when you’re late.

  16. This is definitely the truth. Come to Australia and try walking around Canberra, the capital city: it’s known nationwide as the town of roundabouts. I’m originally from the US, and I’ve driven and walked in Canberra for years. The roundabouts make life easy on the drivers, but on more than one occasion I’ve commented to my partner that the pedestrians turn potential pedestraian areas (e.g., Lonsdale Street, in the CBD) into genuinely dangerous walking areas (dear reader, she agrees with me!).

    They’re great for traffic flow, but absolutely terrible for pedestrians. Rip ’em out.

    1. It depends on the design and the context, the speed limits leading in and out, etc. The one in Berkeley is very sensible, as are many all over Australia. Large swathes of Canberra are pedestrian unfriendly and would mostly remain so with or without circles.

  17. US studies show that roundabouts reduce injuries by 75%. This is an amazing figure. In fact, Wikipedia’s figure is 80%, and 90% for serious injuries.

    How about public support? According to Wikipedia, A 1998 survey of municipalities that built roundabouts found public opinion prior to construction as 68% opposed; afterwards it was 73% in favor. A 2007 survey of citizens found public support ranging from 22% to 44% prior to construction, and several years after it was 57% to 87%

    As for pedestrians, I lived for many years in two countries with lots of roundabouts (Israel and the UK). Roundabouts are not a problem for pedestrians in either of these countries (roundabouts in residential areas are always small).

    1. Having observed roundabouts as implemented in Sydney neighborhoods (as opposed to the UK, Belgium, etc) I think they could work in the US, and if they helped reduce the plague of four-way stops I would be fully supportive of their use.

      However, I think this assertion – “US studies show that roundabouts reduce injuries by 75%. This is an amazing figure. In fact, Wikipedia’s figure is 80%, and 90% for serious injuries.” – has to be qualified with the observation that anything novel works for a while. High-mounted center stoplights on cars, daytime running lights, etc all reduce accidents when first introduced but after they become universal and the novelty wears away so does the effect.


      1. No, it’s actually not a fleeting effect. It’s because virtually all the crashes at roundabouts are sideswipes and shallow-angle collisions, rather than being the much more deadly and/or seriously injurious T-bones and head-ons that happen at square intersections.

  18. There seems to be some confusion regarding circular intersections. DuPont Circle is not a modern roundabout. If you want to see the difference between a traffic circle, a rotary (UK roundabout) and a modern roundabout (UK continental roundabout), search to see pictures. has a video about modern roundabouts that is mostly accurate ( ).
    Modern roundabouts are the safest form of intersection in the world. Visit for FAQs and safety facts.
    The low speed operating environment of a modern roundabout makes them much more efficient and much safer for all users than traffic circles and rotaries.

Comments are closed.