Reason Foundation

Reason Foundation

Public Works Financing

Competing Models for HOT Lanes

Variable-toll, value-priced Model 2 HOT lanes are the road to meaningful congestion reduction

Robert Poole
July 1, 2006

As the person who commissioned the policy paper that introduced HOT lanes to America (in 1993), I'm obviously very pleased with the progress this idea has made, since the 91 Express Lanes first went into operation in California 10 years ago. Today we have HOT lanes in operation on individual freeways in Orange County and San Diego, Houston, Minneapolis, and most recently Denver. Overall, my ever-changing map of "managed lanes" projects (from feasibility studies to PPP proposals to projects being implemented or in operation) now includes 23 urban areas.

But there is a very big difference, in concept and purpose, between two types of managed lanes that both go by the name HOT lane. And before the private sector gets carried away with optimism, companies need to look carefully at which model a proposed project represents.

This disparity was there from the outset in California, with the contrasting approaches of the I-15 HOT lanes in San Diego and the 91 Express Lanes in Orange County. The former are what I call Model 1 HOT lanes. They are conceived and operated first and foremost as HOV lanes, which sell any excess capacity to drivers willing to pay a variable toll. The latter represents Model 2, which is essentially express toll lanes that give special deals to some kinds of high-occupancy vehicles. There is an enormous difference between these two.

Model 1 is championed by planners who want to get people out of their cars, and who generally oppose most additions of highway capacity. Even when they do embrace new capacity (as in San Diego, and so far in Atlanta), Model 1 planners give priority to car-pooling, generally two-person carpools (HOV-2), with revenue-generation a distinctly lower-order goal. Model 2 is championed by those who focus on offering congestion-relief to motorists, and who realize it's going to take a lot of new capacity to do that, paid for by toll revenues.

This is why we see totally different conclusions being drawn about how much of the cost of new HOT lanes can be paid for out of toll revenues. Several studies around the country have concluded that HOT lanes can generate enough to pay for operating costs, but hardly anything for capital costs. That's why San Diego's huge extension and widening of the I-15 HOT lanes is using sales tax and gas-tax funds, not toll revenues. And it's why several HOT lane feasibility studies in Atlanta have reached a similar conclusion.

This has consequences not just for highway finance; it also greatly affects congestion management. This was brought home to me this spring when I reviewed the latest Atlanta study, on the I-75 Northwest Corridor. There, in detailed year-2030 tables showing toll revenues generated during each hour on each segment, I was shocked to see whole blocks of PM peak time with zero revenues. Digging deeper, I found the explanation. Because of the policy priority for HOV-2, during any time period when there were enough two-person carpools to fill up the HOT lanes, they were modeled as being closed to paying customers! In other words, at the peak of the peak�just when you need congestion relief the most--the powerful pricing mechanism is turned off and willing buyers are turned away! (At a recent HOV/HOT workshop, I learned that this same policy is being followed today on the one-year-old HOT lanes on I-394 in Minneapolis.)

In sharp contrast, on the 91 Express Lanes in California, the policy is that HOV-2 vehicles always pay the market price, and HOV-3s get only a 50% discount at peak periods. Thanks to this policy, motorists can count on being able to go 65 mph or better at the worst rush hour. And because value pricing keeps the Express Lanes uncongested (at Level of Service C), during the busiest afternoon peak hours, those lanes handle 49% of the traffic despite having only 33% of the peak-direction lanes. This is the model being followed for the new managed lanes being added to the Katy Freeway in Houston.

What keeps Model 1 HOT lane plans in being is many transportation planners' devotion to reducing vehicle miles traveled (VMT) as their main goal. The idea is that driving is bad (at least driving alone), so policy should encourage ride-sharing, transit, and not traveling. In thrall to this idea, our large urban areas have spent billions over the past 20 years building some 2,500 lane-miles of HOV lane. Yet as HOV lane-miles have soared upward, the fraction of commuters who carpool has steadily declined--on a national level, from a high of 19.7% of commuter trips in 1980 (when we had only 250 lane-miles of HOV), it reached a new low of 10.4% by 2003. And surveys show that in many urban areas, most "car-pools" are actually family members who would likely be traveling together, with our without car-pool lanes. Meanwhile, VMT continues growing, despite the planners' efforts.

With Texas cities leading the way, we are starting to see a change in focus�from reducing VMT to reducing VHT (vehicle hours of travel), i.e., reducing congestion. So the first reason to shift from Model 1 HOT lanes to Model 2 is that the underlying rationale for Model 1 is wrong-headed.

Offering motorists meaningful congestion relief�as with Model 2 HOT lanes�is a very different story. People are desperate for relief, and most are willing to pay for it, at least some of the time. But you can't offer reliable congestion relief if you turn it off just when it's needed most! And more pragmatically, you can't pay for what's needed unless you adopt Model 2, tolling nearly all vehicles under a variable-toll, value-priced approach. Using that approach, the Fluor/Transurban team will pay for 100% of the cost of adding two HOT lanes in each direction on the Washington Beltway. My own calculations (to be published this fall) suggest that the same can be done on much of Atlanta's highly congested freeway system.

Even in metro areas that already have several hundred lane-miles of HOV lane to convert to HOT, a seamless network is still going to be a very costly proposition. It means not just adding new lanes on all those freeway segments that lack HOV lanes; you don't have a seamless network until you add very costly flyover connectors at most of the interchanges. That will never happen using gas or sales-tax revenues; it will only be possible with aggressive use of variable tolling�and ending the absurd free-ride for HOV-2 carpools.

Robert Poole is director of transportation studies at Reason Foundation and has advised the last four presidential administrations on transportation issues. An archive of his work is here and Reason's transportation and tolls research and commentary is here.

Robert Poole is Searle Freedom Trust Transportation Fellow and Director of Transportation Policy

Print This