In many cities, there are carpool or "HOV" lanes which, during rush hour, are only for use by cars with multiple people -- a minimum of 2 or 3. These lanes also exist on bridges and also offer reduced tolls. (Motorcycles and electric cars also get access.)
In many areas, enforcement is lax, and so cheating solo drivers will use the lane, taking the low risk of an expensive ticket, which can range from $300 to $500. They do it to get around heavy traffic and because they don't think they will get caught. Caltrans, which manages California highways, estimated as many as 39% of the cars in the carpool lane may be cheating. When you watch one of those cheaters go by as you are stopped in traffic, you get naturally angry.
Maybe you shouldn't. It turns out that in many situations, carpool cheats may be improving traffic flow and speeding up your commute.
To understand why, you must understand the mistake that was made in original carpool lane design. Carpool lanes are of dubious effectiveness at the goal of reducing congestion and improving the throughput of people on the roads. They can often make traffic worse rather than better. Traffic engineers know this, and in most towns, there are now efforts to change carpool lanes into HOT (High Occupancy Toll) or "Managed" lanes. These lanes act like carpool lanes, but allow solo drivers to pay a toll to use them. The amount of the toll goes up if demand and congestion go up to stop the solo drivers from overflowing the lane.
How can carpool lanes be failing? After all, there are 2 or 3 people in each car in the carpool lane, so that lane is moving more people per hour than the other lanes, even if it's handling fewer cars, as it must to flow smoothly. Indeed, if it's not partly empty it won't offer enough reward to entice carpoolers.
The problem, in most cities, is that the vast bulk of these carpools are natural carpools, which would have existed anyway. A couple heading off to work together. A parent with a child. Very few -- estimates suggest around 10% -- are "induced carpools" where 2 or more drivers said, "hey, in order to make use of the faster lane, let's leave all but one car behind and make a carpool." The induced carpools take a car or two off the road, and are a win. The electric cars don't reduce traffic, but they reduce pollution, so they are a different sort of win. But for most of the cars, all we do is redistribute the existing cars, moving many (but far from all) of the natural carpools from the right lanes to the left.
Lanes in a freeway have a capacity of around 2,000 cars/hour. Try to go over that and they "collapse" into heavy traffic, which reduces the capacity (as well as the speed) and it gets worse until the road opens up again. Sometimes the collapse happens at below full usage, and the lower the usage, the less likely stop-and-go collapse is. Carpool lanes also face a special problem, in that even when they are not being used to full capacity, they sometimes get congested because drivers are afraid of going full speed next to stalled traffic in the next lane. To prevent that, many cities actually try to make their carpool lanes divided from the main traffic, with only a few entries and exits.
In order for a carpool lane to help, it must remove more cars from the road than it reduces the usage of the lane below full. If it removes fewer cars, it can actually be hurting.
Let's do some simple math. Imagine 4 lanes, for a total capacity of 8,000 cars per hour. Rush hour ride pooling is only about 1.2, so that's 10,000 people per hour. Now switch the left lane to HOV-2 (carpool with minimum of 2.)
For the carpool lane to be attractive, it has to run at just under capacity and flow smoothly. Consider a carpool lane moving 80% of full, or 1,600 cars/hour and 3,200 to 4,000 people per hour. Only 10% of those carpools are induced, though, so 160 to 200 cars were removed from the road (there are a small number of 3-driver carpools on an HOV-2.) The road now carries 7,600 vehicles per hour instead of 8,000 in theory, and around 9,300 people instead of 10,000.
If you get the utilization safety up to 90% you get a modest win, but the higher it goes, the more risk of congestion it gets. At lower utilization, though, something far can happen outside the carpool lane. All the non-carpools pushed out of the carpool lane (in order to get it flowing smoothly) go into the 3 regular lanes. This makes it more and more likely they will go into congestion collapse as they are overused. When they do, it drops their capacity, making it hard to recover, and on top of that, the carpool lane often slows too, due to the dislike of having very different speeds.
The carpool lane isn't always making it worse. If you have enough to fill an HOV-3 lane, and a decent number of them are induced, you can get a situation where you improved the road capacity. It's very hard to fill a lane with HOV-3, but HOV-3 plus electric cars can do the job. Casual carpools also do this, but more often than not they are taking people off of transit, not out of cars. But the truth is, it still only works some of the time, and other times it is making it worse. Hardly worth the expense.
When you see a 1/3rd empty carpool lane and slow-moving main lines, things have been made much worse by the presence of that carpool lane. The only way to make that good is to somehow greatly increase the number of induced carpools.
The problem can also be addressed by moving solo drivers into the carpool lane. One popular way to do this is to let green cars go in. It's not just a reward for low pollution, it also helps balance a carpool lane when the demand is not enough to fill it. This works, but it's often not enough, and it's also hard to control. You can't let only the "right amount" of green cars in the lane.
This is where the "managed" or "HOT" lane comes in. You let solo drivers into the lane for a fee. You let in just enough to make best use of the lane capacity, and if too many come in, you raise the price until demand and capacity balance. Everybody wins -- those solo cars reduce the burden on the regular lanes, and the highway gets money to pay for infrastructure.
Carpool cheats
The above numbers are worse when you factor in the large number of carpool cheats. If all those cheats stayed in the main lanes, they would be slowing them down. The cheat gets the boon of the faster lane, and takes the risk of the expensive ticket.
Imagine that cheating were actually blessed. If a cheater does 400 commute trips a year, and the odds of getting a $400 ticket are one in 200, then that actually works out to about $2 of risk each trip, which is not very different from a common HOT toll. This would provide a similar benefit to the HOT lane of moving some solo drivers out of the crowded regular lanes, but it would do so at a vastly, vastly lower cost. Conversion of lanes to HOT is estimated to cost from $2M to $5M per mile, for signs, restriping, tag readers and enforcement. In addition, drivers must all purchase a smart toll tag. This has some similarity to the way many transit systems work, with no gates, just enforcement officers who every so often check tickets.
Of course, having cheats is not as good as HOT toll payers. First of all, tickets for cheating come randomly. Some cheats might go for years without getting caught. Others might get caught three days in a row if they're unlucky.
Tolls can be adjusted in real time to react to demand. Cheaters will see the risk as fairly fixed while the lane is moving well. On the other hand, cheats will naturally respond to congestion and leave the lane if it slows down -- there is no point in taking the risk of a ticket if the lane isn't going to move you significantly faster. If the carpool lane clogs up, the cheats will desert it and open it up again.
It could be possible to do dynamic adjustment of the fine, and have LED signs which show the current fine. It's also possible to do dynamic adjustment of the enforcement to some degree, and to make people aware of that.
One big negative is that enforcement is costly, if you have to send out a cop to pull people over. Costly for the highway, and also costly for the cheat -- you not only pay the fine but get delayed 10 minutes when you were obviously in a hurry.
Enforcement requires pulling people over because today, children count as carpoolers. This is an area of hot political debate, though I am baffled as to why. Carpools with children are very rarely induced and don't take a car off the road. Limiting carpools to people of driving age (or better, two or more car owners) would make enforcement vastly easier, and allow it to even be done with cameras. But try to tell parents they shouldn't count as carpools and watch the dander fly.
(I will also note, before you assume I want to be a carpool cheat, I have never done so personally, and I drive an electric car with unlimited solo carpool lane access.)
Beyond lanes
The original idea of the carpool lane was to offer people a valuable reward to encourage them to carpool. Even improved with HOT lanes (or by cheating) the benefit is still marginal. To really make a dent in road capacity and congestion, you need something much better, which induces more real carpools -- take people who today would drive solo and get them into a carpool. It turns out the special lane is not enough. In future articles, I'll explore new options that might be available in the world where almost everybody carries a smartphone.
