Speed, Mortality, and a Bad Idea

What good is a speed limit here?

Councillor Bill Saundercook of Parkdale-High Park in Toronto wants to lower the speed limit across the city by 10 kph because, he reasons, it will save lives.  As reported by 680 News, Saundercook said: "If you are exceeding 40 kilometres an hour, there's an 80 per cent chance that the accident you get involved in could be fatal. And in fact, the opposite is true the other way. If you are less than 40 [kilometres an hour], then there is an 80 per cent chance that it won't be fatal."

Let's set aside Saudercook's horrendous English.  He meant to say that a pedestrian struck by a car has an 80% chance of dying if the car is travelling over 40 kph, and a 20% chance of dying if the car is travelling under 40 kph.

The Basics

This is a well known statistic: 40 kph seems to be a magic number in terms of pedestrian mortality. However, there are various versions of this chestnut.  In Australia, for example, they say the odds of killing a pedestrian doubles going from 40 kph to 50 kph (there's a video of the effect on crash test dummies too).  Another report includes a nice graph of the correlation between speed and mortality.

The thing is, argues Saundercook, that most people in Toronto seem to drive about 10 kph above the posted limit, figuring that police will not stop them for speeding because 10 kph is a not a particularly large delta vee.  So a limit of 40 kph means cars are really travelling at 50-ish kph, and if the limit were 30 kph then they'd travel at 40-ish kph.  And that should translate into fewer deaths.

There are some big "ifs" here.  For instance, just because people travel 10 kph over the limit in some parts of Toronto doesn't mean they do it everywhere.  And they don't do it all the time.  Exactly how often, statistically anyways, do people speed?  Anything less than 100% of the time will start eating into the number of lives saved by lowering the speed limit, at the expense of inconveniencing every driver in Toronto.

Another "if" is predicting what human drivers will do.  Will they still drive 10 kph faster than the limit?  Or will they now start driving 20 kph faster?  I don't know.  And given how finicky humans behaviour can be, I wouldn't bet on human predictability.  Again, I'm not saying it won't be like the Councillor says it will be, but I'm saying I want more data - and so should we all - before accepting his statements, which are all really just bald assertions.

Lowering the limit from 40 kph to 30 kph is a 25% decrease in speed.  Grade school math informs me that this will add about 25% to one's travel time.  Oddly, the Councillor seemed to disagree with this when asked about it on CBC Radio 1 (sorry, I can't find a link to that conversation).  He claimed that people would get to their destinations just as quickly at 30 kph as they would at 40 kph.  To be fair, when you include stoplights and other interruptions, it may well be that a 10 kph speed decrease wouldn't contribute substantively to travel time - but I'd want to know some hard numbers before buying into it, because my numbers suggest there would be a noticeable difference.

Let's say you have to travel 10 km, and there's a stoplight every 2 km.  Let's say you accelerate and decelerate at the EPA standard rate of 3.78 m/s².  The amount of time and distance travelled while you're accelerating to and decelerating from either 30 kph or 40 kph is negligible (about 2%) compared to the time you spend at a constant speed.  Let's also say stoplights stay red for between 30 seconds and 1 minute.  Under these conditions, travelling 10 km at 30 kph will take between 26% and 29% longer at 30 kph than it would at 40 kph, give or take about 1% to account for the neglected accelerations and decelerations.

This assumes you're travelling in a straight line.  The difference will not be as great if you have to turn at intersections because the amount of time you spend waiting to turn, and actually turning, doesn't depend on your cruising speed.  But turns don't take too long (no matter how long that left turn my seem to take).  I would estimate a 5% contribution due to turns, for the average driver.

This also assumes there's virtually no traffic on the roads.  Traffic slows cars down.  Traffic may slow down cars uniformly to the point where transit time will be determined by traffic and not by speed limits.  But if that's the case, then the speed limit doesn't matter and lowering it won't help.

In other words, I'm just not seeing how the Councillor can claim that drivers will spend as long on the road at 30 kph as at 40 kph.

Saundercook also agreed with the statement by the CBC interviewer that travelling less than 40 kph will save 80% of lives lost to pedestrian deaths.  This is a utter misreading of the data he himself quoted, underscoring a total lack of scientific and mathematical understanding.  Councillor Saundercook would fail any course I ever taught.

The Data

Let's look more closely at the data (what little of it can be found easily).  From articles in The Toronto Star, CityTV.com, a City of Toronto Staff Report, and  the trafficeservicestps blog, I found that:

• In 2007, 22 pedestrians died in traffic accidents.
• In 2006, 20 pedestrians were killed, of which
  • 9 were seniors,
  • 7 were killed at intersections by turning vehicles,
  • 6 were jaywalking, and
  • 6 ran or walked into traffic, trying to catch a bus, walking off a streetcar island, etc.
• More than 2,300 pedestrians are injured every year in traffic collisions.
• In 2009, 15 pedestrians died, a "vast majority" of which involved vehicles turning at intersections.
• About 3 out of every 200,000 Torontonians die each year as pedestrians in traffic accidents.
• SUVs increase the odds of pedestrian death by a factor of between 2 and 4.
• 1.5 - 2 million vehicles travel on Toronto's roads every day.

For the sake of discussion, let's say about 20 pedestrians die each year in Toronto in traffic accidents.  When you consider vehicle usage per year (over 550 million equivalent vehicles per year), we see that the ratio is something like one death in 25 million equivalent vehicles.  Even on a per-person basis, the odds of dying are 0.0015%.  Those are pretty good odds.

Of course, we can do better.  These 20 deaths are still 20 unnecessary deaths.  But do we really need to change everyone's behaviour (by lowering the speed limit) to save these 20 people?  This seems like a highly ineffective way to go about it.

Indeed, what I'm saying is that there are plenty of more effective ways to save at least some of those 20 lives than by just lowering the speed limit.  And we can come up with better solutions by looking more carefully at the numbers.  You'll see that lowering the speed limit just doesn't make sense when you consider the real data.

Better Solutions

We're already well below the 40 kph rule that Councillor Saundercook endorses.  2,300 pedestrians are injured each year in traffic collisions.  Only 20 die.  That's less than 1% mortality, not the 20% we would expect if we all drove 40 kph.

Intersections seem to be a leading site of pedestrian deaths.  I expect that cars turn a corners at speeds  far lower than the speed limit.  That is, the speed limit is irrelevant for those who die because of cars turning corners.  Lowering the speed limit will not help these cases.

Jaywalking seems to be a substantive contributor.  Clearly, lowering the speed limit would save some jaywalkers.  But why should every driver (even those who obey the speed limits) have to slow down because others (the jaywalkers) are breaking the law?  Why not actually have police officers ticket jaywalkers?  I can tell you I have seen thousands of people jaywalking right in front of police officers, and the officers do nothing about it.  The police might reasonably argue that compared to all the other things they have to do, stopping jaywalking is perhaps the least important.  Fine.  Instead, use traffic cameras to identify jaywalkers.  Increase the fines for jaywalking; this will "scare" some people into obeying the law.  Or hire special constables charged with ticketing transgressors for these minor offences that require neither the training, nor the insurance afforded armed police officers.

Or admit that jaywalking is one way to win a Darwin Award, and move on.

Seniors seem to account for nearly half of all pedestrian deaths.  How can we adjust the cityscape to help seniors?  Considering the effect of intersections, perhaps all we have to do is lengthen the duration of walking signals so seniors have more time to cross streets.  I have not been able to find more specific data on why the seniors died; without that information, it's hard to know exactly what the problem is.  But I'd bet it's something that lowering the speed limit does not address.  Remember also that seniors tend to be more frail than younger people.  I would expect the injuries sustained by a senior as a result of a traffic accident would generally be more serious than those sustained by a younger person.  Does that mean we should lower the speed limit by 20 kph?  Of course not!  But I suspect that a 10 kph decrease in the speed limit would not be as beneficial to seniors as Councillor Saundercook thinks.

"Walkers" also constitute a sizable portion of pedestrians who died in traffic.  These are people who were trying to catch buses, stepped off streetcars, etc.  Again, each of these must be treated distinctly.

When people step off streetcars, traffic beside the streetcars is supposed to be stopped.  Speed limits are irrelevant here, because the root cause is not violation of the speed limit, but violation of the law against passing streetcars with doors open.  The two circumstances are not causally related at all.  To make streetcars more safe, add cameras that can record license plate numbers of passing cars, so that the police can ticket them.  Raise hefty fines for drivers who violate the existent laws.  Then watch people get a grip and do the right thing.

Also, we can ensure that every new or renovated streetcar stop has an appropriate island or other traffic barrier, and perhaps even special traffic signals to let people get off the islands and to the sidewalks.

For those who are trying to catch a bus and insist on running out into traffic - well, there's no cure for stupidity.  Accidents will always happen.  We'll never reduce the death rate to zero permanently.  And at some point there will be diminishing returns: to try to save further lives will completely cripple society, which in turn will ruin everything else.

SUVs kill more people than do cars. Obviously, the bigger the car, the more serious the impact it would have on a pedestrian.  (Dropping a paperclip on your head is not the same as dropping a piano on your head.)  We could limit the use of SUVs in cities.  We could lower the speed limits just for SUVs (which might also promote their owners trading them in for more fuel efficient cars).  We could charge a user fee for people who drive SUVs to pay for infrastructure that will make city streets safer for pedestrians.  These are just three ideas off the top of my head; I'm sure a proper design group could come up with many more.

Bottom Line: Saundercook is Wrong

I will say one thing for Saundercook's solution: it's easy to implement from a political and administrative point of view.  That makes it and efficient solution.  Lowering the speed limit means just some paperwork at City Hall, and a bunch of new signs.  (Note, however, that the cost of changing out all the signs could be rather considerable.)  By comparison, the solutions that I've sketched out are much more complex, and so less efficient.

But my solutions are better than Saundercook's because I'm considering the impact on everyone based on far more detailed data than Saudercook seems to care about.  My solutions are more effective than his because they address the problems without necessarily causing even more problems - or at least as many big problems - as Saudercook's solution will cause for everyone who isn't a pedestrian.  They're solutions that balance efficiency and effectiveness; this is what you need for the best possible solutions.  (I've written about effectiveness and efficiency elsewhere.)

So to sum up, Saundercook is wrong: lowering the speed limit not save as many lives as he thinks, and it will be a substantive inconvenience for a huge number of people.  There are plenty of things that can be done to save lives, but we need to design them properly, and that is something that the good Councillor seems unsuited to do.