Sunday, March 23, 2014

Are sidewalks even necessary?

Habitually, one of the first things considered to indicate a walk-friendly area is the presence, or absence, of sidewalks. In general, if none are present, most assume that the area is thus hostile to pedestrians. That was what I thought too... until I chanced upon streets like these in Japan:
Residential street in Tokyo

Another residential street in Tokyo
Yet another one...

Residential street in Nikko
These streets have no sidewalks, there are markings on the pavement that may be seen as delimitations between pedestrians and cars, however, these are treated more like a recommendation than a command. Pedestrians and cyclists both frequently occupy the center of the street. This put into question for me the orthodox thinking that pedestrian-friendly streets must absolutely include sidewalks to ensure safety and comfort while walking. In fact, here is what I have concluded:

Sidewalks exist NOT to make walking more comfortable, but to separate pedestrians from car traffic and thus allow cars to travel at high speeds because they no longer have to contend with other users on  the street.

In fact, when streets are narrow and all users (pedestrians, cyclists, car drivers) are put on the same level, car drivers have no choice, they have to slow down and take care. This can be very frustrating for them, because they have a vehicle that can go very fast, but they are forced to stick to a speed not much higher than a cyclist's.

In terms of safety, it is essential to understand the principle of risk compensation. This principle says that we all have a certain tolerance threshold for risk. When we perceive risk to be too high, we take steps to reduce it. When we perceive it to be low, we adopt behaviors that increase risk. Thus, a street that may seem perfectly safe, for it is wide, has nothing but lawns on each side and pedestrians and cyclists are pushed to the side to bike lanes and sidewalks, may incite drivers to drive faster and to take less care, increasing the risks to all.

On the other hand, a narrow street without shoulders, with fixed objects (trees, buildings, fences, etc...) and where drivers have to watch out for pedestrians and cyclists seems like a riskier environment, so people will drive slower and be more careful.

It results that perception and reality may be different. The wide street is more dangerous for pedestrians crossing it. So oftentimes, the street that APPEARS to be safer is actually LESS safe.

The idea that we have to separate street users at all cost result in streets that are overwhelmingly large. Let's suppose a basic street built on the principle of user separation, here is what it would look like:

A "basic" street to achieve user separation

This street, that separates pedestrians, cyclists, moving cars and parked cars, will thus be 18 meter wide, or 60 feet. That's a minimum. According to the dominant mentality, this is what is required to be safe.

But I think the following is just as safe:
A narrow street
This street is just 5-meter wide and pedestrians, cyclists and cars can all use it. The proximity between users imposes care and awareness of other users. You could even have a few on-street parking spots if you want, this would force cars in both directions to use only one lane when one is parked, but as long as there's not too much traffic, you're fine...

This bears repeating: "as long as there isn't too much traffic". Both streets may be just as safe overall, but the first one has a much higher capacity, and much higher car speeds. So if we're talking of arterial streets or collectors or other high-capacity streets, then yes, user separation makes all the sense in the world to optimize the street capacity, to order traffic and to increase speed. But on residential streets where vehicular traffic is low, the second street is much more appropriate.

Not only would it save the city 60-75% in construction and maintenance costs while permitting higher density, but it will also be a major obstacle to through traffic. No need for traffic-calming measures, the street itself is the best traffic-calming measure there is.

Be warned though, if a residential street is wide, you'll need sidewalks. Pedestrians will naturally stick to shoulders and drivers will perceive the street as safe, despite pedestrian presence on the pavement, since they are far enough away that the driver creates a lane in his mind and follows it. In this case, the higher risk perception will not counterbalance the higher risk from proximity.


  1. "As long as there is not too much traffic"....

    Given your later comments on induced demand really depending on speed, don't you think it's likely that the design of Japanese streets (5-6m) precludes the possibility of too much traffic? In other words, drivers will sense that there is no benefit speed-wise given the constrained road and choose another route. During my four years living in Japan I walked many 6m streets, and very rarely did I encounter a level of traffic that would negatively impact the street's walkability.

    Love your blog - Japan is the reason I got into planning (I wanted to bring their ideas back to BC) so each of your posts make me smile - Japan does so many things right planning-wise.

    1. Bingo.

      At least, as long as there is an arterial close by to offer a quicker route. If there isn't, yet distances are uncomfortably long for walking or biking, it may not suffice to deter driving.

      It seems we come from the same place then, weaboorbanists, unite? I like to say that on my first trip to Japan, I left thinking that "a car is freedom" and I came back thinking "my car is a weight around my ankle". It was a complete perspective shift.

  2. This comment has been removed by the author.

  3. Great post! I've linked to your blog in a recent post of mine, "Narrow streets with people walking in the road"