Gridlock Sam’s Street Safety Fumble in the Daily News

Sam Schwartz is out with a list of street safety recommendations in the Daily News today, which he prefaces with a bizarre warning against lower citywide speed limits:

Today, there’s an outcry to lower speed limits to combat pedestrian fatalities. Wouldn’t it be great if just putting up signs worked? But it probably won’t.

I agree that New York City, not the state, should dictate local speed limits, but I have found no evidence that a blanket change in the city limit from 30 mph to 25 mph or even 20 mph would work.

You may say there’s no harm in just slowing all cars down. But some studies, mostly on highways, have found that this can actually increase crashes. Some drivers lower their speeds; others keep driving as they did. The speed differential between cars increases, and that increases crash potential.

Most readers are probably going to come away from the piece with the notion that 20 mph speed limits are misguided, which is too bad, because Sam’s actual proposals are good (though not sufficient to bring traffic deaths down to a level approaching zero).

The campaign for blanket 20 mph speed limits is about much more than putting up signs. For one thing, it’s getting the word out that speed matters.

Right now, most people don’t even know what the speed limit is on NYC streets. They don’t get why the decision to drive 35 or even 40 mph is an unnecessary risk that, applied to millions of trips every day, results in far more injuries and deaths than if we all stuck to 20 or 25 mph. The campaign for 20 mph can change that. It’s getting a ton of coverage and a lot of attention from elected officials right now. NYC has a window of opportunity to broadly communicate why driving faster than 20 mph impedes reaction times, leading to more crashes, and drastically increases the severity of injuries in the event of a crash.

As for the specific claims about blanket 20 mph speed limits, the speed differential argument that Sam makes may apply to highway conditions, but urban streets are a different beast. In the UK, blanket “signs-only” 20 mph zones are an increasingly popular low-cost first step to improve safety. Research on signs-only slow zones may be scant, but the evidence so far suggests there are real safety benefits, and no studies have shown the type of downside that Sam predicts. Danny Dorling, a professor of geography at Oxford, cites research from Scotland [PDF] finding that “20 mph (32 kph) limits without traffic calming measures at 78 sites found reductions in speed and casualties, concluding that such limits offer a low cost option for promoting road safety.”

The broader point is that a citywide 20 mph speed limit, while not sufficient to achieve drastic traffic fatality reductions on its own, would be a huge step toward making safe driving speeds the social norm. It would reset expectations and raise the bar for engineering and policing. If New Yorkers start thinking of law-abiding driving as adhering to a 20 mph maximum speed, it will be that much clearer that streets need better design and enforcement. Vision Zero is a huge undertaking, and it’s hard to see how it can be accomplished without a significant across-the-board reduction in vehicle speeds on all streets.

I know this wasn’t Gridlock Sam’s intent, but his column is going to muddy what should be a clear message: Speed kills, even speeds most people currently consider acceptable.

  • qrt145

    Increasing the speed differential between motorists increases the risk of crashes between motorists. But given that most of the people killed on the streets of NYC are pedestrians, what we should be talking about is the speed differential between motorists and pedestrians!

  • John Petro

    Let’s not confuse the difference between lowering speeds on highways versus urban streets.

  • Increasing the speed differential between motorists increases the risk of crashes between motorists largely when one motorist is passing another or when one swerves into another lane, as one might experience on a highway. While that’s certainly a possibility on our multi-lane avenues, such a scenario is less likely on single-lane side streets.

    If I’m speeding along at 40 mph and the car up ahead is going 20, the most likely outcome is that I’ll slow down to the speed of the car in front of me. The cultural shift in attitudes about driving speeds among even a small subset of motorists could have a calming effect on lots of other people on the road.

  • qrt145

    That’s exactly the response I give when people say “but no one obeys the speed limit!”. As a matter or fact, some people do, and when they do so they help calm traffic, at least during busy hours. This may not work on an 8-lane highway, but it certainly works on a street.

  • Joe R.

    I’ve said it already elsewhere but I think a nice carrot/stick approach here is raising speed limits on NYC highways where the geometry allows at the same time that we reduce them on local streets. This would have the positive effect of drawing motor traffic on to highways if drivers know they can drive at a comfortable speed without risking a speeding ticket.

  • Keegan Stephan

    Thank you for this piece, Ben. So crucial in the push for Vision Zero.

  • StepUpAndSaySomething

    I don’t think Sam is a dumb guy, maybe there is more to his point he didn’t go into. But until someone makes the point clearly and with data, I’m going to agree with the other commenters, that his speeding message just muddies the water as we’re finally making process. I wish he read this blog; I’d be a huge supporter of his idea if he could show why it’s correct, which he didn’t really do in the article.

    I really didn’t get how he’s saying that lowering the speed limit could make it more dangerous, and then suggests lowering speed limits around schools. Maybe his editor cut something out. Or maybe he’s trying to appease both sides. But with something this important, no one needs appeasement, we need truth and backbone. It’s no longer cars vs. pedestrians; it’s those who want safe streets vs. those who protect the guilty. And fewer and fewer people are on the dark side.

  • ddartley

    Maybe I’m wrong but my understanding is that this “speed differential” notion comes from the 85th Percentile Speed doctrine, which I’m quite sure is exclusively about highways, not city streets.

    I’m really doubting myself here, because Sam Schwartz is much more knowledgeable about this stuff than, say, I am, but I really have a strong feeling that he’s completely off base here. NYC’s internal streets, to my understanding, have little to do with this “speed differential” notion.

  • Who’s to say? Sam Schwartz is a smart guy. It is a difficult question, not so cut-and-dry. Maybe he is right. His opinion certainly deserves some consideration.

  • I don’t see anything in his article saying he’s against lowering the speed limit on single-lane one-way side streets to 20 MPH. He is against a blanket reduction, which would include the multi-lane avenues. Maybe he’s right.

  • Leonard Diamond

    Although there is hope that the lack of enforcement of speed limits might change under the current administration, if things dont change , how does lowering an unenforced speed limit from 30 or 25 to 20 really accomplish anything? Would it be more advantageous for peds and cyclist safety to focus the effort on making NYPD train their officers in making radar stops and in making it a regular part of each precincts efforts?

  • s

    Or maybe he’s not. It might have helped had he cited scientific research beyond non-contextual highway studies. Instead it’s just conjecture that damages the overall effort of safer streets, which Sam is most definitely not against. Sam should be smarter than that.

  • It is an informed opinion based on extensive experience in the field. Maybe the evidence doesn’t exist yet, but his expression of the opinion should lead all of us to question our own beliefs until such evidence can be found.

  • s

    No! It’s a lazy opinion based on the assumption that as an expert he doesn’t need to provide additional research other than “I’ve read some stuff I can’t cite right now.”

    Sam himself says that he’s basing his guess on highway studies, which are wholly different from even NYC’s wide avenues where things that don’t exist on interstates – such as traffic lights, intersections, public transit, pedestrians, commercial and residential buildings, etc. – exist in abundance. And there are plenty of studies on speeds in urban environments from Europe to prove Sam wrong.

    It’s ok. Sam messed up.

  • Bolwerk

    Sam Schwartz is a car guy. Unlike Cuomo, he can be swayed by empirical evidence. He knows that sometimes what works best for cars also works well for transit users. But his priority is still the car, which for the pedestrian is a vehicle of death no matter what speed it’s traveling when it crushes him/her.

  • Bolwerk

    Except the speed on most highways is probably determined more by congestion than by legalities.

  • Joe R.

    Actually, the doctrine says to use the 85th percentile when setting speed limits on local streets, and the 95th percentile for highways (both rounded up to the nearest 5 mph/kph). The rationale in both cases is that the majority of drivers will feel comfortable driving at or below a properly set speed limit. Moreover, most drivers won’t go significantly below the speed limit, so speed differentials will be relatively small. The point of having speed limits is to have a legal means to deal with outliers driving much faster than the average driver. You also have outliers on the other end, such as people who won’t go over 40 mph on a highway. They are arguably just as dangerous as those going way over the speed limit. For this reason, we sometimes also set minimum speeds on limited access highways.

    How does this all apply to NYC’s streets? I’m not really sure it does. Right now the speed limit of 30 mph is already well below the 85th percentile on many streets, particularly arterial streets. In general, a speed limit which is set too low doesn’t get people to slow down without saturation enforcement. We learned this in the mid 1970s with the national legislated 55 mph speed limit. Speed limits should never be legislated. They should be set at the 85th percentile on local streets and the 95th percentile on highways. 75 years of traffic engineering shows this to be the safest way. The problem is the 85th percentile speed on many NYC arterials would end up well over 30 mph, perhaps as high as 60 mph in certain spots. This is obviously not good for vulnerable users. What to do then? Reengineer any streets so the 85th percentile speed falls where you want it to be, whether that’s 20 mph, 25 mph, 30 mph, or higher. Sad to say, there are no shortcuts here. We don’t have the manpower for saturation enforcement of speed limits citiwide. Speed cams at best will only be installed on a small percentage of streets.

    Besides reengineering streets where practical, I highly recommend focusing much more on the one thing which I think is mostly responsible for the type of homicidal driving we often see in this city-namely congestion. When progress is measured in feet at times, people will do all sorts of dangerous things they wouldn’t otherwise do just to gain one car length or make one more traffic signal. If we had far fewer vehicles on the roads, drivers would be much more civilized. There would also be natural gaps in traffic when pedestrians/cyclists could safely cross streets.

  • Joe R.

    That’s true for much of the day but during off-peak hours a rationally set speed limit could divert large numbers of vehicles to highways. Off-peak hours are when pedestrian/cyclist fatalities are at their highest because it’s the time when speeds on local streets are at their highest. If we can get some of these vehicles to choose highways for most of their trip we’ll be far better off.

  • Bolwerk

    I dunno, I’m very skeptical of anything that involves putting more people on highways. I suspect there is less congestion if people distribute across the street system. Compel them onto highways and they quickly overwhelm the artery, and then clog surrounding streets.

    This is why I favor road pricing over tolling arterial routes. Hell, I favor getting rid of arterial routes entirely. It’d be nice if BRT advocates got a clue and started trying to appropriate arterial highways instead of things like the North Shore.

  • Joe R.

    Congestion is an entirely separate issue here which needs to be dealt with even if we keep the status quo on speed limits. It’s a fact there are enough vehicles on the road now to overwhelm both limited access highways AND local streets. That’s best dealt with by a system of congestion pricing. Reduce the number of vehicles to a number which won’t overwhelm the capacity of the highways most times of the day, and you won’t get much spillage into local streets.

    And yes, I think appropriating at least one lane in each direction for BRT is a better use of our highways. Heck, I’m all in favor of appropriating one lane just to make a bidirectional bike highway. Yes, you’ll need separate access ramps, but some means for bikes to travel rapidly/safely around the city is something we sorely need anyway, and the highway already exists. Better to use it for BRT and bikes than mostly single occupancy motor vehicles.

  • Bolwerk

    Certainly there are too many cars. But the point is that way fewer cars can cause that congestion when we lean heavily on arterials, and it will>/b> overflow to local streets. It’s counterintuitive, but if you think about it arterials are pretty low-capacity – something car people can’t grasp. At higher speeds, they probably mean considerably lower overall throughput.

  • Bolwerk



  • ddartley

    Great. I just spent 10 minutes writing a reply and then I lost it.

    Hopefully I’ll get a chance to do it again. In the meantime, let me just quote a definition, for anyone else reading who might not know, of what the 8th pctl principle is: “This
    is the speed at or below which 85% of the traffic moves. For example, if 85 of each 100 motor vehicles
    were recorded at 45mph or under, then 45mph is the 85th percentile speed.” (

    I just lost about four paragraphs explaining why I find that
    inappropriate for urban streets (both with AND without newly applied traffic calming), which maybe I’ll get a chance to re-write some time today, but I think its inappropriateness for urban streets kind of speaks for itself anyway…!

  • StepUpAndSaySomething

    I’m happy to consider and even support Sam’s viewpoint, but it doesn’t make a lot of sense to me and goes against the data people are talking about. I’m hoping that Sam goes to it in more detail and with backup. We’re all on the same team, willing to do what it takes to decrease street violence.

    Someone mentioned this journal, which said in part:

    “The relationship between vehicle speed and accident outcome severity is well established. A major study conducted by the OECD and the ECMT in 2006 concluded that speeding is the number one road safety problem in most countries around the world, and that reductions in average speeds of approximately 5 per cent would yield a reduction in fatalities by as much as 20 per cent (OECD/ECMT, 2006). Research also indicates that even modest speed reductions can prevent the occurrence of collisions and significantly reduce the outcomes of those crashes that do occur; particularly those that involve vulnerable road-users who are more predominant in the urban environment (e.g. Kloeden, et al., 1997; 2001; Elvik et al., 2004; Nilsson, 200”

    It goes on to talk about the other benefits of capping traffic speeds, like decreasing pollution and noise. And it mentions that limiting the top speed doesn’t increase travel time much, since that’s driven more on weather, drivers mood, and infrastructure. It’s a good read, check it out:

  • Joe R.

    Whether or not an 85th percentile speed limit is inappropriate for urban streets is moot. It’s a fact that under free-flowing conditions (which often don’t exist for much of the day in NYC) traffic finds its own speed, regardless of the speed limit, and interestingly even if there is no speed limit. To ignore this fact is like ignoring the fact that water flows downhill when engineering a sewer system. The traffic engineer has a great many tools in their toolbox to lower the 85th percentile speed to a value appropriate for urban settings. You’ll get no argument from me that people drive too fast on local streets. That’s a fact, and it must be dealt with. I suppose in a perfect world you could have speed cams everywhere to enforce any arbitrary speed limit but that’s not the world we live in. There’s a lot we can do now at low cost, including narrowing lanes, eliminating lanes, installing bollards between traffic lanes at intersections (this has the nice effect of both slowing vehicles and preventing the obnoxious jockeying for position near intersections), installing speed humps or chicanes, and having uncontrolled intersections (yes, those keep speeds under ~20 mph because motorists driving faster risk colliding with cross traffic).

    By the way, I often copy my posts as I’m writing them and paste them in a text file (which I save every few minutes) in case Disqus has problems. This has the bonus of giving me a nice, lengthy text file full of my posts which comes in handy. Often I can cut and paste part of a previous post when a similar topic comes up. That really helps because I have pretty severe carpal tunnel syndrome which gets aggravated if I spend too much time on the computer.

  • Joe R.

    Yes, arterials are indeed pretty low capacity per lane compared to highways. Typical traffic light timing cuts the potential capacity right in half.

    Anyway, when I talk traffic reduction, I’m thinking of 90% in the long term. That’s where we need to be to realize Vision Zero. It’s also a number which could be nicely handled mostly by expressways.

  • Joe R.

    One thing which bears mentioning here is if you engineer urban roads for travel speeds of ~20 mph or less, you can pretty much do away with one of the major sources of delay-traffic signals/stop signs. In practice this means you might get where you’re going faster with a 20 mph speed limit than a 30 or 40 mph speed limit where half the time you’re stopped at red lights. Moreover, yet get less variability in travel times by getting rid of traffic signals. That’s much more important. Most people would rather have a travel time of 30 minutes plus or minus 2 minutes than 28 minutes plus or minus 15 minutes. In the former case, you only have to leave 32 minutes early to allow for a near worst case scenario. In the latter you must leave 43 minutes early, but on average you’ll arrive 15 minutes ahead of time. Those 15 minutes are often wasted time, especially if start times of appointments are inflexible.

    By the way, most cyclists are living proof that speed limit reduction often results in minor or no increases in travel time. Cyclists seldom can travel at urban speed limits, and yet in many cases their travel times are similar to automobile travel times, sometimes even less. All motorists do by traveling at 40 mph is get to the next choke point faster. As mentioned in the document you linked to, aggressive driving in cities rarely results in significant reductions in travel time.

  • Bolwerk

    Yeah, I can see that, but expressways are just such blight I’d mostly rather see them dispensed with. Modern els can at least have modest footprints. There is no way to do that with a highway.

  • ddartley

    I admit I don’t get outside of Manhattan (CBD and below) more than a few times a month, but in my experience of real-life conditions on local streets in that limited area, 30mph is NOT below the 85th; indeed it’s well above it. Every once in a while, one or a very few vehicles might get up to 30mph for one block or sometimes a few blocks, but never more than that. So even by the wisdom of traffic engineering (another thing I criticized in my @#$% lost comment), 30mph is too high a speed limit for the roads I see; I guess (not sarcasm here) we’re observing different roads.

  • Joe R.

    Yes, that’s the issue. You’re coming from a Manhattan perspective. The times I go to Manhattan it’s rare I see vehicles going near the speed limit. However, remember the key is to observe traffic during free-flowing conditions. This almost never occurs during Manhattan outside of late nights. The 85th percentile under free-flowing conditions is what traffic engineers use to set the speed limit. Arguably, the speed limit is moot in Manhattan during much of the day due to congestion, so it doesn’t matter if it’s 20 mph, 30 mph, or even 60 mph. I’ll even go so far as to say speed probably doesn’t factor into pedestrian injuries/deaths all that much in Manhattan outside of off-peak hours. From my perspective, it seems like failure to yield is a big problem. It also seems like large trucks with poor visability cause a large amount of carnage, even traveling at low speeds.

    The outer boroughs are a different animal. In many cases traffic is free-flowing much of the day. The 85th percentile speed on many arterials is well above 30 mph due to street design which resembles highways. Speed is a major factor in pedestrian deaths. This is where my set of proposed solutions is likely to be effective. In Manhattan I think heavy enforcement of failure-to-yield, mirrors or cameras for blind spots on large vehicles, and congestion pricing are the most appropriate measures. Speed is mainly only an issue during off-peak times but there aren’t all that many pedestrians around then. Regardless, in the long run we should take measures to reduce the off-peak 85th percentile speed to 30 mph or less.

  • No wonder your posts are always so long and wordy! 🙂

  • Daniel

    Maybe we can find some studies? Or convince a university or the DOT to get a study done? If there is opposition among reasonable people to lowering the speed limit to 20 mph I think it behoves us to counter those that hunch that it won’t work with some data.

    I’d like to just get the state to pass a bill allowing
    the city DOT to set the speed limits on our roads and then the DOT and
    the NYPD can run some experiments to see under what conditions lower
    speed limits help.

  • QueensWatcher

    I have to agree with Joe R. on this. Drive down Queens Blvd, Northern Blvd, Woodhaven, Blvd, Atlantic Avenue, etc., and the 85th% is easily above 40mph. with no enforcement and wide open roads that encourage faster speeds there is nothing slowing these drivers down. Changing the speed limit is almost just symbolic without intense enforcement and changes to street design. That is why we need to significantly redesign these streets so that the street design, more so than a speed limit, compels drivers to drive at slower speeds in a more rational and predictable manner.

  • chekpeds

    Neck downs and speed bumps are a more efficient way to actually reduce the speed without costly and improbable human intervention


How London Is Saving Lives With 20 MPH Zones

One of London’s 20 mph zones, with physical traffic calming measures and the speed limit prominently displayed. Image: ITDP-Europe via Flickr. When Mayor Bloomberg announced that the new pedestrian spaces in Midtown are here to stay, he made special note of the safety improvements on Broadway, which he called "reason enough to make this permanent." […]

Vacca Endorses Life-Saving 20 MPH Speed Limit

Speed kills, even when traffic is moving at New York’s citywide limit of 30 miles per hour. According to the UK Department for Transport, if a driver hits a pedestrian at 30 miles per hour, the victim only has a 55 percent chance of surviving. At 20 mph, the pedestrian has a 95 percent chance […]