Regardless of the reason for these findings, the study I think does show one thing clearly: to reduce pedestrian fatalities to near zero, a city has around two options:
|The Modernist solution: total pedestrian segregation.|
- Remove pedestrians from the city to the extent possible, or segregate them.
- Remove cars from the city, or greatly constrain them.
By contrast, in dense urban areas where large numbers of pedestrians and high-speed traffic share the streets, it appears a certain rate of pedestrian deaths is inevitable regardless of the measures which are taken to redesign street layouts. Presumably this is due to induced pedestrian traffic: as an urban environment becomes more pleasant for pedestrians, pedestrian activity will rise in response, resulting in a numerical increase in fatalities even as the per capita fatality rate falls. If the new pedestrian volume does not come at the expense of more dangerous automobile use (e.g. if people drive to the pedestrian-friendly area), the number of total transit-related fatalities may remain essentially the same. Possibly this explains why cities of such varying pedestrian friendliness have roughly similar fatality rates: New York (1.9/100,000); Houston (2.0); Los Angeles (2.0); San Francisco (1.6); Atlanta (1.6); Philadelphia (1.7); Denver (1.7).
|Running the gauntlet, 3rd Avenue, NYC.|
Jacobs envisioned a long-term strategy in which pedestrians would, bit by bit, encroach upon and reclaim the public realm from the car. This is quite different from a modern Complete Streets approach, which at best seeks equality for pedestrians, rather than privilege, and which at worst, as Strong Towns Blog has shown, results in over-engineered, car-friendly thoroughfares where sidewalks and bike lanes are little more than an expensive afterthought. To fully address street safety, the conception of the role of the car in urban settings could stand to take a bit more inspiration from Jacobs' point of view.
See also: Are Narrow Streets A Realistic Objective?
(h/t Pedestrian Observations).