Gelinas Responds to Komanoff on Bike-Share Safety

Manhattan Institute fellow Nicole Gelinas submitted this response to Charles Komanoff’s critique of her weekend opinion piece in the New York Post about bike-share safety.

Charles Komanoff, in his Streetsblog post, called my weekend Post piece on bike-share “intellectually muddled.” In the piece, he offers no evidence for any intellectual muck on my part. Instead, he uses mirage to distract from a very real safety challenge. Novice bicyclists need to learn that truckers are not their friends. The way to get that message across to new Citibikers is to tell them, as often and simply as possible.

Let’s go through Komanoff’s piece point by point (without making any broad, unsupported generalizations about it, as he did with mine).

I am not a “bike-share detractor.” I favor bike-share, something that anyone who asks would know. The Post article in no way disparaged the bike-share concept. I challenge Komanoff to find a disparaging statement anywhere in the piece. I have ridden the Velibs in Paris for years. My personal opinion on this should not matter to anyone but me, so I feel no need to trumpet it in print. But if someone makes a factual reporting error in mischaracterizing my approach to bike-share, as Komanoff did, he should correct it.

Komanoff characterizes my pointing out that “Three people died in Paris’ first year of bike share” and that “New York should heed Paris’s lesson” as a “ghoulish lede.” Ghost stories are ghoulish. My facts are simply facts. Three people indeed did die during the first year of Paris bike-share. This is a simple stated fact, a fact I haven’t seen reported elsewhere in the New York press. Further, it is hard to argue that we shouldn’t heed lessons that we can glean from any useful data.

Is the Paris data useful? To be useful, Paris’ data would have to be relevant to New York. Komanoff implies we should ignore Paris in favor of cities such as Boston, Washington, and London, three cities that haven’t experienced bike-share deaths. Whether we should compare New York to Paris is where Komanoff presumably could offer some reasonable statistical debate. Yet he does not.

As I wrote in my Post piece, there is reason to believe we can make the comparison to Paris. Boston and Washington come nowhere close to New York when it comes to population density. Core London, too, is not as dense as New York. Paris, on the other hand, does come close to New York, with 80 percent of our population density. As New York’s own Department of Transportation said in a bikeshare study, “the population density of New York’s medium- and high-density areas – including parts of Brooklyn, where bikeshare will be — is virtually identical to Paris.”

Paris’ bike-share, which started off with 6,000 bikes, 500 more than New York’s will, is off the charts compared to bike-share usage in other cities. Compare its 27.5 million first-year rides to London’s six million, or to Washington’s million, or to Boston’s half-million.

It is difficult to know how many riders New York bike-share will attract in its first year, because the city hasn’t produced any estimates. But even before bike-share launches, New York has now signed up more annual members – 10,000 – than the city of Boston has attracted in two years.

When we have no idea how many people will use our bike-share and when we can reasonably compare our system to Paris’ because of density, it is statistically supportable – although not ironclad – to conservatively estimate, for safety purposes, that our system will get tremendous usage and that we can extrapolate lessons from other cities that enjoy tremendous usage for our own benefit.

If Komanoff has some reason why it is not statistically appropriate to compare Paris to New York, then he should say so.

Paris is a particularly potent example for New York, because – something that didn’t fit in the Post piece – Paris’ pre-bike-share experience in cutting traffic deaths was similar to New York’s in recent years. In 2001, Paris had 114 traffic deaths across all modes (see chart). Because of great public outcry, Paris spent the early 2000s doing the same type of traffic-calming measure New York has done in recent years. In 2006, Paris’ death total was 64. Adjusted for population, that would be 239 deaths in New York, lower than New York’s 274 deaths last year. Paris had an even lower bicycle death number in 2006, the year before it started bike-share, with 2 deaths, compared to 5 in 2001, despite big increases in bicycle ridership before bike-share started there. In New York, by contrast, the number of people killed in bike crashes was 18 last year and 18 in 2006.

That means that Paris started off bike-share in a safer position than New York does, both in terms of absolute numbers and in terms of getting the bicycle-death numbers (not the rate) down in recent years. This is not a good sign for New York.

If there is some reason we should ignore this data, Komanoff should say so. Komanoff is a statistical expert, but his main argument is that we should use his own figures rather than third-party city and state data, as I did, to estimate New York bicycling population.

There’s another reason why we should pay particular attention to Paris’ deaths. First, statistically speaking, bike-share in Paris doubled bicyclists’ deaths. During the three years before bike share, the average number of people who died annually on a bike in Paris was 2.67. During the three years after bikeshare, the average number killed annually rose to 5.67. All of this increase was due to bikeshare.

We can also learn qualitative information from these deaths that can save lives in New York. For example, the first three people killed in Paris bike-share were all women. This fits with the theory that novice bikers, statistically women and older people, are at risk. Second, all were killed by a large truck or bus (two trucks, one bus). Indeed, virtually all of the eight bike-share deaths in Paris over nearly six years (the latest death last October) have involved a large truck making a turn. The majority of the victims have been women, or older people, or both, neither category of which fits the statistical profile of the middle-aged, male expert biker. It is reasonable to assume that inexperienced cyclists have made up the preponderance of victims.

Given these facts – and they are facts – it would be wise for Citibike to print right on bicycle handlebars a warning to new bikers to watch out for turning trucks. Indeed, Paris bike-share has offered these handlebar instructions since the first few deaths, and the number of deaths has gone down significantly.

Komanoff, though, is oddly against such a warning, literally “mock[ing]” it. Of course, the city should also force truckers to follow the rules of the road, as he suggests. But why not do both? Why not arm bicyclists with all of the information they need to stay alive?

Such a warning is especially necessary because the number of speeding tickets the NYPD gave out last year declined 7 percent. Komanoff may want the city to police truckers, but the city is not doing it. That does not mean that bicyclists should die in the meantime. The fact that Citibikers won’t be “as battle-tested as the current biking base” is an argument in favor of such a warning, not against one.

Komanoff also criticizes me for “segment[ing] dangers to bicyclists from general road dangers facing everyone.” This is factually incorrect. I said early on in the piece that “More cycling can make riders collectively safer, by increasing awareness by drivers.”

I also wrote that “New cyclists should understand that bicycling is collectively beneficial” (collectively generally means everyone). Moreover, I wrote that “since the early [Velib] deaths,” as Paris has adjusted its behavior, “all bicyclist deaths have fallen to just more than two annually.”

How do these factual statements not address the traffic-calming measures and overall beneficial impact that Komanoff mentions? (Something that didn’t fit in my Post article, but is interesting: since Paris started bikeshare, all traffic deaths have fallen dramatically, from an average of 56 annually in the three years before bikeshare to an average of 44 now. Injuries are down, as well.)

Finally, it is surprising that Komanoff makes two unforced statistical errors by implying that just because Velib victims would have been safer had they stuck to their pre-Velib way of getting around, the same thing is not true in New York. “Don’t tell that to NYC pedestrians,” he says.

But unless he can give us a measure of how many pedestrians are on a given sidewalk at any given time and then give us a death rate using that denominator of the total population at risk, both for Paris and New York, he cannot possibly make such an assertion. Anecdotally, far more people walk on the sidewalks of Manhattan (or elsewhere) than bike.

Last year, bicyclists made up 6.5 percent of people killed in New York traffic crashes, far more than their 1 percent share of people coming into Manhattan via traditional commuting modes, including the most common ways of commuting, subway and train rides. With bike stations concentrated in Manhattan, many bike-share riders will be switching from subways and buses, or from walking. Subway riders make up more than half of commuters into Manhattan, but only 10 percent of deaths (not including subway suicides). Even using Komanoff’s method of counting bicyclists, cyclist fatalities outweigh their share of the commuting population.

Statistically, you are undeniably safer on the subway than on a bike. This information is useful to you if you are thinking of switching from the subway to a Citibike. This does not mean that you shouldn’t make the switch. But being armed with facts helps you by making you realize the statistical risk you are taking.

To recap: For my article, I spent a week studying archival Parisian news articles from a variety of sources as well as official state and city data from France to learn about Paris’ cycling fatalities, talked to two bicycling advocates in France (one off the record and one on), talked (on background) to city officials involved in bikeshare to ask about bikeshare safety measures, and attempted to contact a leading New York bike advocate to see what his group is doing on bike-share safety education.

To rebut my article, Komanoff wrote a lazy piece that wrongly reported the basic facts laid out in my article. He added nothing of value from his statistical expertise either to rebut, criticize, or buttress the statistics in my piece.

After all that, he essentially acknowledges that I am right. “Traffic hazards await bike-sharers,” he says.

Yet he criticizes “tabloid titillation”!