The speed of cities


Jason Collins


January 10, 2011

Over the weekend, I listened to a great Radiolab podcast in which Bob Levine was interviewed about the pace of walking in cities. Bob spoke about how people tend to walk faster in larger cities, with this relationship surprisingly consistent. Where does this walking pace comes from. As the host Jad asked, do we make the city, or does the city make us?

The early movers in this area of research were Bornstein and Bornstein, who between 1972 and 1974 went to 15 countries across Europe, North America and Asia and measured the speed of pedestrians. They took a 50 feet stretch in similar downtown areas of each city and measured the speed of single, unencumbered walkers traversing that distance.

The slowest walkers were from Itea, Greece (population 2,500), who took an average of 22 seconds to cover the 50 feet. In Prague, a city of over 1 million, the pedestrians covered the distance in a flying average of 8.5 seconds. Walking speed and the log of the population were strongly correlated (with a correlation coefficient of 0.91 that was significant beyond the 0.001 confidence level). Particularly surprising is the consistency of the results and the absence of any large outliers. For example, the five largest cities sampled all had higher average walking speeds than the slowest five.

This high level of consistency raises some obvious questions. While there were no severe outliers in the sample of 15 cities included in the paper, are there any cities that are different? If so, why? Also, where there is variation in the sample, can this be explained? Some other researchers have considered this since the Bornstein and Bornstein paper was published, and I hope to post about those in the near future.

Bornstein and Bornstein based their explanation for the consistency on the total number of people in the city as opposed to density. Borrowing from Milgram, they considered that a higher number of people in a city causes increased stimulation. A person seeking to control the sensory overload will have a higher walking speed as a protection from this excessive environmental stimulation. The variance in walking speeds between cities is the manifestation of this adaptation.

While this is an interesting and possibly correct explanation, two other effects strike me as relevant. The first is the selection effect. Why do people choose to live in cities? If people who have a preference for a faster paced lifestyle wish to live in cities, and those who prefer a slower pace tend to leave, this will tend to drive the results we see. There is an element of the people making the city. This may tend to reduce any “stimulation effect” as residents who live in the city may be there as they wish to be stimulated. The easiest way to reduce stimulation would be to simply leave the city.

The second effect concerns the opportunity cost of time. For the pedestrians, what is the value of an earlier arrival? If incomes are higher in the city, as they tend to be in larger cities, their time has higher monetary value. If there is more or better entertainment in the city, there is more value in being at the destination than dawdling along the way.