Thursday, 9 October 2014

Fair Warning and Photo Radar

     I've written before about speed limits, and generally argued against speeding, so it may come as a surprise that I have received seven photo radar tickets. Admittedly, six of them came over several months after I had my license plate stolen by some nitwit who apparently drove a 1985 Supra, if the images on the summonses are to be trusted. I had reported the plate stolen promptly, so I never had to pay any of the six, and needless to say, it wasn't me speeding.
     The seventh was just this past year, and it was for going 60 km/h in a 50 zone, which wouldn't have bothered me but for the fact that it was a short stretch of the road between two zones where the limit was 60. Also, it did surprise me a little, because I had always understood that there was an unwritten rule that they wouldn't issue a ticket if you were within 10 klicks of the limit, not that I have ever thought that was a valid legal argument against any actual speeding ticket. And also it was probably my wife driving.

     Now photo radar has become a hot topic again here in Edmonton, with the mayor responding on his blog to a petition of angry drivers who want photo radar abolished. They argue that it is a cash cow, that it doesn't actually make us safer, and that speed limits are too low anyway. I'll not address those again here, but instead, I wanted to consider what Mayor Iveson said (and I've said in the past myself) about speed limits, and how exceeding them at all is illegal, period, end of story. I agree with that, of course, but I want to argue here for why there ought to be a buffer as a matter of policy, and how it should be handled.

     First, the reasons for a buffer. I've read that highway engineers usually try to establish speed limits based on the 85th percentile of free traffic flow, meaning the speed at or below which 85% of vehicles are travelling. Although vehicles may vary in their performance and individual drivers vary in their skill and tolerance of risk, taken in the aggregate they can give a pretty fair idea of what people can handle safely and comfortably. I imagine they probably set the actual limit a little bit below that point (perhaps they just round down to the nearest 10 km/h), which would make sense because traffic flows most efficiently when everyone is going close to the same speed, and if only 15% of drivers feel comfortable at or above the posted limit, you'll likely have a lot of people going considerably slower and gumming up the works. So the optimum limit should be something a clear majority of drivers can confidently handle.
     But there's a curious fact about posted limits, which grocery store owners understand. Put up a sign that says "Limit 5 per customer", and people who normally would only have bought one will buy four more. To some extent, the same psychology applies to speed limits, and so drivers who might otherwise have been content at 48 km/h will feel they're missing out on something of value if they don't snap up those extra 12 klicks. So even if they do set the limit at the 85th percentile, it seems likely that the general flow of traffic will usually be at or near the speed limit, which is, after all, a good thing: we want everybody to be going approximately the same speed.
     Now, posted limits are one thing, but the facts of driving are such that sometimes you need to adjust your speed upward or downward in order to make certain maneuvers, such as getting into position to change lanes. Ideally, to avoid exceeding the limit, you'd just slow down and drop back behind the car next to you so you can change lanes, but in practice that's not aways the safest or best choice (especially given the prevalence of tailgating). So occasionally going a few klicks over the posted limit is a perfectly reasonable thing to do for certain maneuvers, and ought not to be discouraged when it's done responsibly in that kind of context. Moreover, I'd be willing to bet that traffic engineers who prescribe a speed limit are building into it assumptions that normal traffic flow will include such minor incidental variations around that value.
     As well, there's the scarcity of attention; a driver has only so much of it, and we want drivers to focus their attention where it is needed most. Although they shouldn't completely ignore the speedometer, micromanaging it is not a good investment of attention, either. A too-rigidly enforced speed limit without any buffer will begin to punish drivers for the wrong thing: watching the road. You could argue that to avoid this problem, you can just set your speed around 5 klicks below the posted limit to leave yourself some wiggle room, and of course that's true. However, remember the "Limit 5 per customer" phenomenon, and the fact that traffic engineers almost certainly take this into account when setting limits in the first place. It's likely that when they post a limit of 80 km/h, they expect and intend for traffic to comply by driving at 80 km/h and not 75 km/h.

     And yet, speed limits are legal limits; if you exceed them, you are breaking the law. If you build into it a formal buffer of 10 km/h, then in effect you're really just raising the "actual" speed limit by that amount. So what to do?

     One of the advantages to a live traffic cop pulling you over is that he or she has some discretion to let you go with a warning, when a warning is sufficient and effective. Why not have photo radar do the same thing? When it catches you exceeding the limit by 10 km/h or less, it would send you not a summons, but just a warning that you've been caught speeding. There would be no penalty, but if you get into the habit (that is, if you get too many warnings within a reasonable time period), you will start being fined.
     Mayor Iveson explained, in his blog, that the proceeds from photo radar do not go into general revenue but are used to fund traffic safety initiatives, so it seems to me that this would fall right within that mandate. The infrastructure for processing such cautions is already in place, so this would be a cost-effective way of delivering a message to exactly the people who need to hear it.


  1. You are correct that recommended speed limits are usually recommended as 85% of desired speed. So if the average speed of free flow travel is 105 km/h, then recommended speed would be approximately 90 km/h. This is regardless of weather any cars are uncomfortable driving below that speed. There's no statistical spread consideration on which drivers are more or less comfortable on what speed, just the average desired speed. This speed is calculated based on road factors (how wide the lanes are, how big the shoulders are, how far away light standards are, etc.) and not studied for any individual road. This stuff is mostly considered for highway speeds.

    The posting of any speed limit, however, is a political choice. Most communities have speed limits for certain types of roads (50km/h for local roads, 60 km/h for access roads, etc.) regardless of the design speeds of the roads. Some engineering these days goes into reducing the design speed for roads by narrowing them, having parking lanes, speed bumps, etc. so that desired speed matches the local political choice in road speeds. Even if the 85th percent desired speed of a road in Alberta was 130 km/h, the provincial government still wouldn't post it above 110.

    I say this because pursuing any comment on the engineered speed or design speed of the road is not a valid point of debate. Your local or provincial government sets the speed, and the road is designed accordingly. Even if it is designed for faster travel than the local government allows, it's the politics which need to change to actually change the posted limits.

    Once you get to that point in the discussion, any talk of buffers and whatnot is irrelevant. There's a limit, follow it or risk breaking the law and suffering the penalty for it. It's a moral choice.

    I would also argue that it's a moral choice for politicians setting the limits to ignore whatever desired speed that the road has and enforce a lower limit. This will tempt more people into breaking speed limits. Governments are not, in my opinion, moral entities.

  2. I will defer to your expertise, although I'm not sure we've understood each other correctly with regard to the 85 figure. I understood from my reading that number to mean the 85th percentILE of traffic, not 85% of some "desired speed", for which the question "desired by whom?" arises.

    I agree that ultimately the setting of speed limits is a political question. However, so too is the question of policy on how to administer those limits.

    Whenever you're making rules, the key question is always: "How do we want people to behave in response to this legislation?" In the case of speed limits, we can take a very narrow approach to this question and just say we want people to drive slower. But I think we should also look at the overall policy objectives as well, which in this case means traffic safety and the efficiency of our roadways generally. It wouldn't be helpful, after all, to adopt enforcement strategies to bring speed down if those strategies created bigger problems elsewhere. And, as the principal objective of speed limits is to maximize the efficient movement of traffic, simply slowing everyone down isn't necessarily a desirable objective anyway.

    My suggestion for a buffer doesn't need to be anchored with the speed limit at the bottom. We COULD have photo radar issue warnings to people when they are clocked at any point within 5 km/h of the speed limit. "Hey, you came really close to speeding there, dude!", which could be effective if we decided, politically, that we wanted traffic speeds to average somewhere under, rather than at, the posted limit.

    Personally, I don't think that's very practical, because in my experience a posted limit of 60 km/h means most people will be doing their best to go at least that fast, and usually a little over. The intuition is a natural one: if it says 60, I can go 60, so don't hassle me for going over 55. So it'd be a difficult intuition to dislodge, which is why I think it would be more useful to have the warning buffer kick in AT the limit rather than just below it.