Why You'll Love Paying for Roads That Used to Be Free, Part Two


Eric A. Morris is a researcher at U.C.L.A.’s Institute of Transportation Studies, concentrating on a variety of transportation issues including history, economics, and management. Earlier this week, he wrote the first half of a provocative essay on road tolls. Here is the second half.

Why You’ll Love Paying for Roads That Used to Be Free, Part Two
By Eric A. Morris
A Guest Post

In my prior post, I blogged about introducing variable tolls on America’s highways. The basic idea: fight congestion by imposing tolls that vary in response to traffic levels. When roads are too crowded, hike up the tolls, keep some drivers out, and thus keep traffic free flowing at all times.

This idea is getting a lot of traction with opinion-makers and transportation officials, but a skeptical public has yet to be convinced. There is good reason for this; several perfectly valid arguments can be raised against tolling.

Is it fair that government will be charging for roads that motorists have already paid for through the gas tax? Won’t this policy benefit the wealthy, who can easily afford the tolls, and punish the poor? And what will happen to the drivers who avoid the tolled facilities? Won’t conditions in the lanes that remain free degenerate, as refugees from the tolled lanes pour in?

These are good questions, but toll advocates have equally good answers. Believe it or not, converting some of the lanes on a congested freeway to toll lanes should benefit everyone, even the people who choose to never use the tolled lanes.

The reason has to do with the curious mechanics of traffic congestion. When few cars are using a road, speeds are high, but the light volume means few cars get through. Add more cars and eventually speeds start to slow, but the increase in volume means that throughput rises. When a road is just crowded enough so that speeds are around 45 m.p.h., the most cars are pumped through the system.

But add even more cars and trouble starts. Speeds break down, taking throughput down with them. When roads are severely congested, you get a paradoxical situation: the more cars you jam in at one end, the fewer come out the other end.

By pricing to keep traffic speeds at 45 m.p.h. or a bit higher, the toll lanes will work with maximum efficiency. They’ll move a lot more cars through than they did when they were congested. During the peak periods on SR 91, the toll lanes handle 40 percent of the traffic despite the fact that they constitute only one-third of the road surface. So the toll lanes will actually ease the burden on the free lanes, hence the benefit even to those who never choose to pay.

What about the argument that only the rich will be able to afford these new “Lexus lanes”? Ed Sullivan of the California Polytechnic State University has extensively studied the express lanes on California’s SR 91, America’s first variable toll facility.

He finds the equity issue isn’t nearly as clear-cut as it may seem. Those with higher incomes and education do use the toll facility more than others — but not that much more. Many low-income people use the facility frequently; you don’t have to be rich for your time to be valuable. Moreover, many wealthy drivers do not use the facility at all. In all, Sullivan feels the “Lexus lanes” argument is largely a red herring.

Interestingly, the most important factor in predicting who will use toll lanes is not class, but gender. Women use the SR 91 tolled lanes 10 percent more than men do. This is probably due to the fact that women’s travel patterns are more complicated than men’s, and that even in this enlightened era, women undertake a disproportionate number of the household-serving trips. Since women have a more difficult time balancing home and work responsibilities, they are more likely to take advantage of the time savings. In a weird way, tolling may strike a blow for gender equality too.

Still, isn’t this just an excuse for government to get its hands on even more of your precious dollars? Uh, well, O.K., you got me on that one. Let’s not pretend government is jumping on this bandwagon due to fear of the political might of the transportation economist lobby. Although transportation officials are certainly cognizant of the many benefits tolling will bring, fundamentally this issue is getting political traction (six states have operational pilot programs) because the revenue is desperately needed.

Transportation agencies at all levels of government are in a serious financial bind. Thanks to the waning purchasing power of the fuel taxes (again, see this), they are facing unfunded maintenance backlogs, and fancy new projects seem like pipe dreams. Tolls are a very promising source of new revenue.

But is this a bad thing? Our transportation system is in trouble and tolls are a fair way of raising the revenue to maintain it. Shouldn’t users of the transportation system bear the burden of its upkeep?

Even better, paying government to use the roads would get us something for nothing. When you pay a toll, the money is transferred from one party (you) to another (the government). Granted, it is annoying to be the one doing the paying, but at least the money goes to a (presumably) good cause, such as an improved transportation system.

But when you sit stuck in traffic, your time is wasted and no one is benefiting. Better to transfer money from one pocket to another than to let all that time go up in smoke.

And there are other benefits. Since travelers will want to split the cost of the tolls, they’ll have an incentive to rideshare. And new possibilities will open up for transit, since buses that use the tolled lanes will now be able to provide high-speed express services, even on the most congested routes.

O.K., O.K., congestion pricing won’t cure the common cold or bring peace to the Middle East; but it’s rare that a public policy can produce so many winners with relatively small costs.

Selling variable tolling to the voters will be an uphill struggle, particularly when it comes to adding tolls to previously free lanes. There is considerable skepticism about this policy. But there is hope.

The early projects show that motorists initially have doubts, but they become enthusiastic converts when they see and use the facilities. According to the last survey, over 70 percent of SR 91 express-lane users — and even over half of the nonusers — approve of the use of variable tolls. My guess is that someday you will too. Here’s hoping it will cost you $6 to drive home in the near future.


Wait... presumably right now, 33% of the cars travel on 33% of the road space. So if your argument is that with 40% in the Lexus Lane, which is still 33% of the road, there's more room in the free lanes, you're correct. But why are people paying to be in the toll lanes, if it's now even more crowded than the lanes were before the tolls began?


I love these abstract discussions of adding tolls to highways that have been previously been toll-free.

But has anyone actually tried to discuss this with people who know about road design? It's one thing to design a road with tolls in mind. It's quite another to add tolls to a road that was not designed with tolls in mind. The latter type of road typically has *far* more frequent opportunities for going an an alternate non-toll route than the toll road.

I think, for example, of the rather breezy thoughts in the Boston area of adding tolls to I-93. If it were done close to Boston, however, the very design of the road network with I-93 would deeply frustrate the goal.

For all their apparent realism, these discussions of congestion pricing seem very airy when it comes to fitting them with the actual design of road networks.

This airiness indicates to me that the idea has yet to be taken seriously. When someone finally deals with the gritty issues of road network design, that might be a sign of seriousness.


Bobby G

As a member of the skeptical public, I also am yet to be convinced.

Encouraging rideshare is the only solution proposed that reduces the actual count of vehicles that need to get from point A to point B, and most likely it will be a weak effect. If at a free level there are already too many cars to hit optimal speeds, creating a price should remove cars, right? Sure. But those are people that will not be using that method to get where they're going. One cannot say that the burden will be lessened on both the toll lanes and the free lanes without either reducing the DEMAND for cars to get to work every day or shifting traffic to surface streets, which seems to me to be an inevitable consequence.

Likely you've also thought of this problem, but I haven't seen a response I like yet (yes I'm picky): what happens to people who want to use the highway multiple times per day? Now they will be punished... again encouraging surface street usage over a larger distance? Will the price be so low during non rush hour that people will not care? One-purchase per day won't work... some people might camp early-morning light times to lock in their cheaper price for the day... granted this might be corrected by prices increasing as more people start to camp, but now it seems like an unintended, non-equilibrium effect of the program.

Not that I'm necessarily a fan of giving our government more money, I still assert that traffic is not a market inefficiency in and of itself... it's an externality to a successful and popular gasoline market. Reducing traffic and gasoline consumption would both be positive effects of an increase in the gas tax... and even if levels of both didn't improve, the government would make more money.

I know personally I would prefer paying a gas tax more than paying tolls on a freeway, even if I ended up paying twice as much annually. The incentives are more solid with the gas tax I think, it's much more difficult to subvert, and the less effective it is at accomplishing its goals (of less consumption and less traffic), the more money would be made for the government.



Mike, the point is that the "curious mechanics of traffic congestion" mean that you can NOT equate throughput with road space. The point of variable tolls is to use the toll price to restrict the number of cars on the toll lanes to the one that allows maximum throughput, i.e. somewhat below the "crowded" level. Ideally, the toll lanes have free-flowing traffic and can manage far more throughput on the same space than the non-toll lanes. Actually, if you applied the same restrictions to the entire road, it would allow more cars to pass then when it's congested - you'd get to you destination quicker if you had to sit still at the entrance to the road until it's your turn to enter at the maximum-throughput traffic level, instead of trying to get on as quickly as possible and spending most of the time on the road moving at snail's pace.

David Heigham

There ain't no such thing as a free freeway. We drivers and citizens pay one way or another for their maintenance and improvement. The options are tolls where people who want the extra convenience pay; gasoline, etc., taxes where all drivers pay; or other taxes where all citizens, drivers or not, pay.

Tolls are something that I don't like; even the modern ones which avoid having to stop at toll booths. I don't like them even when they are voluntary, and I choose to pay for the convenience. But they are better than the alternatives.


@Karl: Let's take I-93 as an example. So far as I understand, the main proposal was for a barrier toll at the NH line. That would (of course) lead to extra traffic on Route 28, but it's not such an awful idea and it's separate.

But if you were going to create a 'Lexus lane' on 93 north of town between 128 (I-95 to outsiders) and the Big Dig, that's not so ridiculous. Make it a 'high-speed lane to Boston only' inbound and a 'high-speed lane to Woburn only' outbound (no intermediate exits), take over the HOV lane where it exists and the passing lane where it doesn't, maybe create an underpass northbound so that traffic coming off the Sumner can access it, make it Fast Lane/EZ-Pass only, and sort out the north end in conjunction with the rebuilding of that awful cloverleaf. It can be a reasonably inexpensive add-on to the cloverleaf work.

Nobody's saying it costs nothing (although, say, on I-84 in East Hartford it _would_ cost close to nothing to convert the HOV lane) but the cost is worthwhile.



it's my understanding that politically, every time a retrofit from non-toll road into toll road comes about, the appearance of unfairness in the public's eyes has caused enough of an outcry to cause them to be closed down immediately. Not enough political backbone to make them work.

Also, there are the public-private toll partnerships mentioned occasionally, where the toll money goes NOT into a 'transportation improvement' general fund, but instead into the pockets of the developers. Those last just seem skeezy to me.



The current discussion is about the NH border so that it is politically less costly, and with fewer environmental impact issues. But past discussions have discussed placing toll booths in the Somerville/Medford stretch. Won't work due to the road network there.

James A

"Even better, paying government to use the roads would get us something for nothing. When you pay a toll, the money is transferred from one party (you) to another (the government). Granted, it is annoying to be the one doing the paying, but at least the money goes to a (presumably) good cause, such as an improved transportation system."

You've got to be kidding me. Something for nothing? And from the government no less? It's amazing how one paragraph can destroy any credibility you had in this writeup. All should know that the gov't is one of the most inefficient users of capital the world has ever seen. Once they get a taste of new tax revenue, it won't be just used for roads. It will then be used to plug "revenue gaps" in the general fund or be "diverted" to other "essential" needs. Anyone heard of the Educational Lottery? How much is going for education and how much is plugging gaps in other funds? The idea you propose is interesting, but to put gov't in such grandeur is an insult to intelligence.


Ed Kay

And don't forget the safety issue. For years the State of Illinois brazenly proclaimed that its toll roads were built to be safer than its free roads. Here's the ditty broadcast endlessly on the radio, sung to chorus of Loch Lomond:

You'll take the highway and I'll take the tollway,
and I'll get there safely before you.


Road design is an optimality problem.

This article is merely a persuasion piece to encourage people to even bother with the problem.

Thus, I think appealing to the road design issues at this point doesn't discredit the idea. It is a matter of theoretical vs. practical concerns.


I echo Karl's sentiments - I think that tolled lanes are obviously better suited to particular roads simply because of the design and amount of space available. The amount of capital that would have to be thrown into making certain roads toll-ready, though a long-term investment, is probably staggering enough for the project to be pushed aside for quite some time, at least in Ohio. But I can think of roads like OH-315 as it passes through the city, and even I-270 in Columbus that might be suited.


wouldn't the use of variable tolls require new roads /lanes to put them on? And wouldn't the additional roads / lanes decrease congestion? You can't retrofit all lanes with tolls because you would still have capacity issues where suface roads are not an option.


In response to Mike, I think the idea is that if you were to drive those lanes today without the congestion pricing, the amount of traffic the lanes experience is variable. So, while the average load in the lane may be 33%, with so many more cars the actual travel time may vary considerably depending on what time of day it is. The congestion pricing would allow for users to be assured that they will always be allowed to travel at a minimum speed of 45 miles per hour, which isn't something they get normally. I think that's what they would be paying for.


It's not so much a question of adding tolls, in some cases, but varying them based on traffic conditions.

In Stockholm they're seeing success in reducing both traffic congestion and emissions through a congestion-charging system.


Such a system may not be "fair" to everyone, but it is not inordinately unfair.

Full disclosure: I work for the company that helped implement this system.


I understand congestion control but beleive we need to seperate the cost component from the argument. We already have transporation tax dollars being spent on non-transportation projects so advocating yet another tax is a long row to hoe.

Lets assume the argument is true that restricting traffic entering into a freeway to hold average speeds at 45MPH or above actually INCREASES throughput. If that is the case then lets just do it with timed and gated on-ramps. Everyone gets the same hypothetical benefit without added taxes. I'm also assuming these gates are cheaper than building a new road...

The authors argument also assumes all of the current toll information is available to the travelinng public the instant the routing decision is made. More reason for public resistance... a constantly changing toll cost that will change in a way that will be perceived as arbitrary.

Now, what I'd love to hear are opinions on what the actual unintended consequences of interstate congestion pricing will be :-)


Mike B

Once again I have to ask why are you putting so much effort arguing for HOT lanes. I don't know any of reasonable party that is seriously against the use of HOT lanes. Even HOV -2 lanes are underutilized, increasing highway costs are big with the green lobby, are preferred over increasing gas taxes by the conservative lobby and by targeting those with disposable income it is popular with the liberal lobby.

You seem to be picking a fight with uninformed BANANAs, which is a fight that's not worth having. Even Pennsylvania is likely to slap toll booths on the currently free I-80 over its entire length, so point of view has already prevailed.

I believe the problem that people have isn't with tolls (or even variable tolls) on premium travel routes (State turnpikes when built were basically nothing but large "Lexus highways"), but the wide area congestion charging zones sticks a nip down to the grocery store with a $10 "convenience charge". Not only do the overhead costs of such systems virtually wipe out the profits, they remove even the option for those with low opportunity costs to pay with time rather than cash.

As an effective "road monopoly" the state must regulate itself to provide consumers with a range of driving options, some premium, some basic, and allow for a range of payment options in terms of cash and time. Geographic mobility is a basic necessity that when made unavailable traps people in disadvantaged circumstances.



The systems design on this one intrigues me.

1) How to efficiently inform drivers of current toll rates and travel delays, as well as alternative routes available. Optimal route choices depend on knowing this before you arrive at a tollbooth.

2) Construction and transactions costs. Modifying roadways and building tollbooths is expensive - when will the project pay for itself? What new costs are introduced to driving? (e.g. time through tolls, exact change or ez-pass devices, distraction of selecting efficient route)


All roads should be private. Tolls, although the inevitable outcome of a free society, will not help the immoral and inefficient "public" highway system. Anyone with experience dealing in government highway contracting understands the reality of the situation.


There is something a little disheartening about the idea of erecting gates and concrete barriers in the middle of an interstate as a way of increasing the flow of traffic.

I find the entire concept of toll roads to be misguided. Toll roads put a barrier on various forms of commerce, not just in the cost of the toll, but in the collection of the toll which often costs businesses more lost revenue than the cost of the toll itself. In other words, the cost of _paying_ the toll is actually, in terms of opportunity costs, more than the cost of the toll. This lost revenue also represents a loss of tax revenue.

We increasingly have the means of making real time traffic monitoring data available to drivers. As this technology matures and becomes mass market, I believe that many of the traffic problems that we have will be able to self-correct much quicker as people's route changes dynamically in response to congestion and accidents.