Various comments I've heard and read since that post have motivated me to revisit this matter. Not that I've changed my mind; indeed, if anything, I am even more convinced that the approach being taken by Toronto's government is entirely inadequate.
I wrote recently about some of the problems with Toronto's new and lower speed limit on certain residential streets.
While there's no question that lower speeds save lives and lessen the odds of serious injury, there are three assumptions built into that argument: that vehicle speed is the sole culprit, that one universal solution is adequate, and that there is a singular, linear relationship between lowering the speed limit and death and injury rates that is entirely decoupled from all other factors.
Unfortunately, none of these assumptions hold.
Vehicle speed is certainly a proximal cause of the energy of impact, which is the cause of injury and death. But equally proximal are other circumstances of an accident.
For instance, some accidents happen because pedestrians jaywalk, or because they're too busy with their smartphones to notice an oncoming vehicle. Lowering the speed limit will punish all drivers (by lengthening their commute, making them be away from work and family more, and so on) for the faults of a few pedestrians. That’s just not right.
In the case of jaywalking, a better solution is to stop the jaywalking such that neither drivers nor non-jaywalking pedestrians will be inconvenienced or harmed. How exactly to stop jaywalking is a separate problem, and will depend on circumstance. Perhaps physical barriers will work in some places, dedicated pedestrian paths in others. But just as different solutions will work to address jaywalking, so too must different solutions be developed to lower vehicle speeds. One size simply does not fit all.
Some accidents involve bicycles. Lowering the speed limit might lower the rate of bicycle/car accidents, but it won’t do anything for bicycle/pedestrian accidents. Road sharing by both bicycles and cars is a well known stressor for both riders of bikes and drivers of cars, especially in North America. Increased stress is a source of risk, so managing speed limits is little more than a band-aid solution unless one also addresses the stress. The definitive answer in this regard is, I think, to stop sharing the road - that is, to develop true bicycle lanes, as they’ve done in Copenhagen. Clearly, this is a costly solution, but the long-term benefits are gargantuan in terms of increased safety, decreased pollution and fossil fuel usage, increased general health due to exercise, and general community well-being. And since bicyclists are more likely than car drivers to notice and stop at shops on their route, one can expect economic benefits as well.
No matter what solutions we decide on, there will always be consequences, rippling effects that can spread quite far. We're quickly learning, thanks to systems modeling, how to predict those consequences. Unfortunately, systems modeling is largely unknown to politicians and the general public, so everyone keeps thinking that simplistic solutions like lowering speed limits is the best we can do. Still, at least some of the systemic consequences arising from some interventions can be easily identified, if not quantified, without any special training. All we need to do is invest the time to think it through.
I’ve already mentioned some of the more obvious consequences to lowering the speed limit in my last post. Here, I’ll touch on some consequences of alternatives to lowering the speed limit (remembering that the ultimate goal - and the basic design question - is not to slow down cars, but to decrease risk of injury).
One way to do this is with speed bumps. There are, however, two significant and problematic consequences of speed bumps. First, emergency vehicles, snow-plows, and other municipal vehicles will have a hard time getting through an area with speed bumps. Second, drivers behave more erratically in the presence of speed bumps; they repeatedly speed up between bumps and then slow down. This increases both fuel consumption and risk of accidents.
Another way is to make lanes narrower or to clutter the roadway with distractions like medians filled with brightly coloured flowers. This is well known to slow down drivers. Related to this is curbside parking, which also causes drivers to slow down for similar reasons. But do we really want to introduce even more distractions, knowing as we do that distracted drivers have more accidents?
Alternatively, we could go the other way and radically decrease distractions by removing all stoplights, traffic signs, road lane markers, etc. Without those guides, drivers have no cues as to what they should expect, and will therefore be more cautious. This may sound like a recipe for disaster, but it seems to work very well in Drachten, which clearly demonstrates how an entire community benefits in many diverse ways from a radical design intervention to decrease injuries on roads.
Public transit is another potential solution. Even if we just consider buses, we can immediately see benefits. One bus can conservatively replace 20-30 cars. Congestion would plummet, decreasing the stress felt by car drivers, and speeding up transit times. In combination with other measures designed to encourage lower speeds, the overall decrease in vehicles on the road cannot but lower the risk of injury. Street cars on dedicated lanes and subways would decrease the risk even further, but it wouldn’t really help matters in residential/suburban areas, where, it seems, so many accidents happen.
And then, we could get really creative.
For instance, since the skill of a driver is a factor in assessing risk, and since some drivers can drive faster than others and still be safe, perhaps we need to get rid of absolute speed limits and instead use relative measures. That would mean coming up with a graded scale, perhaps based on icons and colours (colours alone would be bad for colour-blind drivers). There might be three types of roads: high-traffic residential/urban streets and rural roads, major urban/suburban arteries, and highways/freeways. Armed with that information, each driver can judge how fast they may safely drive.
Note that this immediately takes into account a fairly common practice by traffic police of allowing some “slack” when they are checking for speeders. There’s a bit of a practice, for instance, in the Toronto area, to not stop a driver going less than 10 kph over the speed limit. While some people consider this an urban legend, I know that such practices exist. Given this, the absolute speed limit is already turned into more of a suggestion than an obligation. Why not enshrine that practice more deeply in how we manage traffic?
Obviously, this would not affect drivers’ responsibilities to adapt to road conditions (rain, snow, etc.), visibility (day/night, fog, etc.), and other factors. But these are already generally accepted responsibilities, so they “factor out” of the equation, as it were.
By removing the absoluteness of speed limits, one introduces a little uncertainty, and that uncertainty would likely cause drivers to err on the side of caution - i.e., travel a bit slower than they otherwise might - as well as pay more attention to their surroundings.
One objection to this idea that comes to mind immediately is that such a scheme could well increase the difference in speeds among different drivers. The cautious, elderly woman driving her grandchildren home from school in her minivan may end up going slower than she would otherwise, while the twenty-something male “stallion” in his sports car may end up driving faster than he otherwise would. This difference in speeds among drivers can translate into an increased risk. More detailed analysis would be needed to determine if overall risk would increase or decrease in such a scheme.
Finally, we must consider what the near future holds. It seems inevitable that self-driving cars will be coming to a roadway near you. I think it’s only a matter of years now, especially given the overall stellar record of devices like the Google Car. At that point, one can completely restructure the system by which speeds are set, and all of this hewing and crying about lowering speed limits may have been for naught. The advantages of self-driving cars from environmental, traffic, health, and safety perspectives, are huge. (They’d be even more spectacular if we had self-driving public transit!) If we set aside the whole question of pedestrians for a moment, it’s easy to imagine a road system without any stoplights or signals at all because the self-driving cars will simply coordinate their own movements amongst themselves.
Of course, such solutions are years off, but if that’s how things are going - and I do think that that’s how they’re going - then shouldn’t we be planning for that and not obsessing about short-term measures that will likely only be relevant for a few years?
So I hope I’ve made it clear here, that there are many alternatives open to us besides the politically expedient yet unproven solution of lowering speed limits. The question is: Do we have the balls to actually do the work necessary to find a good systems-based solution, or will we just sit back that screw the pooch?
 |
| Source: The Toronto Sun. |
Unfortunately, none of these assumptions hold.
Vehicle speed is certainly a proximal cause of the energy of impact, which is the cause of injury and death. But equally proximal are other circumstances of an accident.
For instance, some accidents happen because pedestrians jaywalk, or because they're too busy with their smartphones to notice an oncoming vehicle. Lowering the speed limit will punish all drivers (by lengthening their commute, making them be away from work and family more, and so on) for the faults of a few pedestrians. That’s just not right.
In the case of jaywalking, a better solution is to stop the jaywalking such that neither drivers nor non-jaywalking pedestrians will be inconvenienced or harmed. How exactly to stop jaywalking is a separate problem, and will depend on circumstance. Perhaps physical barriers will work in some places, dedicated pedestrian paths in others. But just as different solutions will work to address jaywalking, so too must different solutions be developed to lower vehicle speeds. One size simply does not fit all.
Some accidents involve bicycles. Lowering the speed limit might lower the rate of bicycle/car accidents, but it won’t do anything for bicycle/pedestrian accidents. Road sharing by both bicycles and cars is a well known stressor for both riders of bikes and drivers of cars, especially in North America. Increased stress is a source of risk, so managing speed limits is little more than a band-aid solution unless one also addresses the stress. The definitive answer in this regard is, I think, to stop sharing the road - that is, to develop true bicycle lanes, as they’ve done in Copenhagen. Clearly, this is a costly solution, but the long-term benefits are gargantuan in terms of increased safety, decreased pollution and fossil fuel usage, increased general health due to exercise, and general community well-being. And since bicyclists are more likely than car drivers to notice and stop at shops on their route, one can expect economic benefits as well.
No matter what solutions we decide on, there will always be consequences, rippling effects that can spread quite far. We're quickly learning, thanks to systems modeling, how to predict those consequences. Unfortunately, systems modeling is largely unknown to politicians and the general public, so everyone keeps thinking that simplistic solutions like lowering speed limits is the best we can do. Still, at least some of the systemic consequences arising from some interventions can be easily identified, if not quantified, without any special training. All we need to do is invest the time to think it through.
I’ve already mentioned some of the more obvious consequences to lowering the speed limit in my last post. Here, I’ll touch on some consequences of alternatives to lowering the speed limit (remembering that the ultimate goal - and the basic design question - is not to slow down cars, but to decrease risk of injury).
One way to do this is with speed bumps. There are, however, two significant and problematic consequences of speed bumps. First, emergency vehicles, snow-plows, and other municipal vehicles will have a hard time getting through an area with speed bumps. Second, drivers behave more erratically in the presence of speed bumps; they repeatedly speed up between bumps and then slow down. This increases both fuel consumption and risk of accidents.
Another way is to make lanes narrower or to clutter the roadway with distractions like medians filled with brightly coloured flowers. This is well known to slow down drivers. Related to this is curbside parking, which also causes drivers to slow down for similar reasons. But do we really want to introduce even more distractions, knowing as we do that distracted drivers have more accidents?
Alternatively, we could go the other way and radically decrease distractions by removing all stoplights, traffic signs, road lane markers, etc. Without those guides, drivers have no cues as to what they should expect, and will therefore be more cautious. This may sound like a recipe for disaster, but it seems to work very well in Drachten, which clearly demonstrates how an entire community benefits in many diverse ways from a radical design intervention to decrease injuries on roads.
Public transit is another potential solution. Even if we just consider buses, we can immediately see benefits. One bus can conservatively replace 20-30 cars. Congestion would plummet, decreasing the stress felt by car drivers, and speeding up transit times. In combination with other measures designed to encourage lower speeds, the overall decrease in vehicles on the road cannot but lower the risk of injury. Street cars on dedicated lanes and subways would decrease the risk even further, but it wouldn’t really help matters in residential/suburban areas, where, it seems, so many accidents happen.
And then, we could get really creative.
For instance, since the skill of a driver is a factor in assessing risk, and since some drivers can drive faster than others and still be safe, perhaps we need to get rid of absolute speed limits and instead use relative measures. That would mean coming up with a graded scale, perhaps based on icons and colours (colours alone would be bad for colour-blind drivers). There might be three types of roads: high-traffic residential/urban streets and rural roads, major urban/suburban arteries, and highways/freeways. Armed with that information, each driver can judge how fast they may safely drive.
Note that this immediately takes into account a fairly common practice by traffic police of allowing some “slack” when they are checking for speeders. There’s a bit of a practice, for instance, in the Toronto area, to not stop a driver going less than 10 kph over the speed limit. While some people consider this an urban legend, I know that such practices exist. Given this, the absolute speed limit is already turned into more of a suggestion than an obligation. Why not enshrine that practice more deeply in how we manage traffic?
Obviously, this would not affect drivers’ responsibilities to adapt to road conditions (rain, snow, etc.), visibility (day/night, fog, etc.), and other factors. But these are already generally accepted responsibilities, so they “factor out” of the equation, as it were.
By removing the absoluteness of speed limits, one introduces a little uncertainty, and that uncertainty would likely cause drivers to err on the side of caution - i.e., travel a bit slower than they otherwise might - as well as pay more attention to their surroundings.
One objection to this idea that comes to mind immediately is that such a scheme could well increase the difference in speeds among different drivers. The cautious, elderly woman driving her grandchildren home from school in her minivan may end up going slower than she would otherwise, while the twenty-something male “stallion” in his sports car may end up driving faster than he otherwise would. This difference in speeds among drivers can translate into an increased risk. More detailed analysis would be needed to determine if overall risk would increase or decrease in such a scheme.
Finally, we must consider what the near future holds. It seems inevitable that self-driving cars will be coming to a roadway near you. I think it’s only a matter of years now, especially given the overall stellar record of devices like the Google Car. At that point, one can completely restructure the system by which speeds are set, and all of this hewing and crying about lowering speed limits may have been for naught. The advantages of self-driving cars from environmental, traffic, health, and safety perspectives, are huge. (They’d be even more spectacular if we had self-driving public transit!) If we set aside the whole question of pedestrians for a moment, it’s easy to imagine a road system without any stoplights or signals at all because the self-driving cars will simply coordinate their own movements amongst themselves.
Of course, such solutions are years off, but if that’s how things are going - and I do think that that’s how they’re going - then shouldn’t we be planning for that and not obsessing about short-term measures that will likely only be relevant for a few years?
So I hope I’ve made it clear here, that there are many alternatives open to us besides the politically expedient yet unproven solution of lowering speed limits. The question is: Do we have the balls to actually do the work necessary to find a good systems-based solution, or will we just sit back that screw the pooch?
COMMENTS