An NYC Citi Bike. One rider says it weighs about 45 lbs and ennabled her to average 9.7 mph on the flat

There’s widespread acknowledgement that there’re remarkably few casualties associated with bike share schemes. Boston’s Hubway system had three reported crashes after 700,000 trips (only one serious) and Nice Ride in Minneapolis had four reported crashes after 600,000 trips (none serious).

Three riders on New York’s extraordinarily successful new Citi Bike scheme experienced minor injuries after 500,000 trips (none serious). This study concluded that larger North American bike schemes (> 1,000 bikes) incur an average of just 4.3 injuries per year.

What’s especially interesting is that bike share schemes are reported to be much safer than regular cycling. For example, there were 14.6 million cycling trips in Washington DC in 2011 and 538 reported bicycle crashes. After 4 million trips, the city’s Capital Bikeshare scheme had half as many reported crashes per trip. One estimate for New York is Citi Bike users crashed at a rate 2.5 times lower than regular cyclists. Bikeshare’s safer in Europe and other countries too.

Bike share tariffs usually incentivise shorter trips so comparing crashes on the basis of trip rates might give a misleading answer. Perhaps bike share isn’t any safer than ordinary cycling after all. I don’t have any data to come down either way on that question so for the purposes of this exercise I’m going to accept the conventional wisdom that bike share’s a lot safer than regular cycling.

The intriguing question then is why it’s safer. I haven’t found any studies that formally investigate this question so I’m going to advance some speculative explanations, mostly drawn from what I’ve heard and read (these are framed relative to regular cycling):

One: the sorts of bicycles used in bike share schemes are safer than standard bicycles. They’re heavier and have wider tyres, so they’re slower. They’ve usually only got three gears so riders are less likely to ride in hilly terrain. They’re more upright too, which arguably gives more control and makes the rider more visible to drivers. They usually have lights built-in so they can be seen at night.

Two: drivers are more careful around bike share riders. Since bike share schemes are still relatively new, motorists are cautious about this new phenomenon (this might not last). Or it might be drivers assume bike share riders are more likely to be inexperienced so they take more care around them. Perhaps some feel greater empathy with bike share riders because they’re usually dressed in standard clothes, not lycra.

Three: bike share users are accountable via their credit card for any damage to the bicycle. They accordingly ride with greater care.

Four: bike share schemes operate in the most cycling-friendly parts of cities. These tend to be in and around the CBD where cycling infrastructure is usually better. Congestion, high pedestrian density and controlled intersections often mean traffic moves slowly. In contrast, casualty statistics for regular cycling usually cover a much larger area, much of it with rudimentary provision for cycling.

Five: bike share riders are different. They’re more likely to be drawn from those who don’t own a bicycle and therefore have limited experience with cycling. Rather than their lower skills exposing them to greater danger, it might make them more cautious. They might, for example, intuitively select the safest routes and eschew those that look problematic due to traffic. They might also tend to be drawn from a demographic (e.g. older) that’s better able to judge risk than cyclists in general.

While I can see all these factors contributing to varying degrees, the fact is there’s no definitive evidence. My instinct nevertheless is the first three are the least important. I expect the last two are the ones that really matter, especially number five – the profile of bike share users tends to differ from that of regular riders. For example, Stephen Miller at Streetsblog notes:

In DC, for instance, women comprise 45 percent of Capital Bikeshare’s membership, but only 23 percent of the general cycling population. In addition, 70 percent of Capital Bikeshare members did not already own a bicycle before joining.

If that proposition is right, I’d expect to see much less difference in casualty rates between bike share and regular cycling in places where cycling has a very high mode share, like some European and Chinese cities.

Whatever the explanation, it’s important to bear in mind that bike share still accounts for a small fraction of cycling trips even in cities with long-established schemes (a relative term – there were only seven schemes worldwide in 2002, now there’re almost 500). Policy-makers need to understand that what’s true for bike share users might not hold in all cases for the much larger population of regular cyclists, and vice versa.