I've long been skeptical about the concept of shadow tolling, in which the state goes out to bid for a concessionaire to design, finance, build, and operate a transportation project but the project does not charge tolls. Instead, its financing is based on a long-term agreement by which the state makes regular payments based on traffic counts (and/or other performance measures) over the long term of the concession agreement.
My concerns have been several. First, since the overwhelming reason why governments are interested in long-term concessions is a lack of funding, shadow tolling is a bust because it adds no funding to the pot. To be sure, it's a financing technique, but just like GARVEE bonds, it relies on the state committing to use a portion of its future transportation revenues (mostly fuel taxes). As states that have issued a lot of GARVEE bonds are discovering to their dismay, those future revenues will cover a lot less of future operating and maintenance costs, given what needs to be spent first on servicing this debt. Real tolls, by contrast, add large new amounts of funding to the highway system.
Second, shadow tolling necessarily means foregoing the huge advantages of value pricing as a tool for managing traffic flow to optimize the performance (speed, throughput, and reliability) of the priced roadway. Since the majority of what we need to invest in added highway capacity over the next several decades will be for urban congestion relief, it makes no sense to add such capacity without value pricing.
I could see a limited set of cases where, despite these drawbacks, shadow tolling could make sense. For projects intended to divert truck traffic from congested "free" roads (like the planned Miami Port Tunnel), charging a toll would discourage at least some of the intended traffic from using the facility. And there can be other public policy reasons (apart from political fear of tolling) to use shadow tolling instead of real tolling: new roadways needed for hurricane evacuation but lacking sufficient daily traffic flow to fund the bulk of the project, or El Paso's project to accommodate an expansion of the Fort Bliss military base.
But two recent articles have led me to see a somewhat wider role for a specific form of shadow tolling. Michael Parker's column in PWF's January 2007 issue and Patrick DeCorla-Souza's paper in Public Works Management & Policy (January 2006) both suggest models that combines value pricing and shadow tolling. As the latter explains it, "Although the private partner would set the real toll rates to manage demand and ensure that traffic is free-flowing, all toll revenue would go to the public sector, and the public agency would reimburse the private partner with a flat fee for each vehicle served at free-flow speeds." He calls this approach "concurrent real and shadow tolling."
In what circumstances might this approach be preferable to real-toll concessioning? Parker suggests that some types of HOT or TOT (truck only toll) lanes might be good candidates, in cases where the traffic is difficult to predict and hence the project is difficult to finance. In the TOT lanes arena, for example, we have yet to see any projects financed, partly because the idea is still new but also partly because forecasting truck traffic on such lanes seems to be far more difficult than traditional traffic & revenue forecasting. Financiers tend to want usage of TOT lanes by trucks to be mandatory, but that approach is unacceptable to the trucking industry. And if the TOT lane concept includes the enticement of allowing double- and triple-trailer rigs (currently not allowed on most Interstates, but strongly desired by many truckers and shippers), the uncertainties in traffic forecasting get even more complex—what rate of investment in such new rigs do you assume, and on what basis?
Likewise, colleagues doing traffic & revenue studies of proposed HOT or Express Toll lane projects tell me that these forecasts are far more sensitive to small changes in assumptions than are the forecasts for conventional toll roads. In other words, there's a significantly higher risk that the forecast will be wrong, especially in the early years. That, too, makes financing difficult.
But if the concurrent real and shadow toll approach were used for such projects, and the concessionaire were paid partly based on availability and partly a negotiated amount per vehicle, it would be able to adjust the real toll rate to be lower if traffic was slower to use the toll lanes than expected (or higher if it was unexpectedly higher than forecast). The pricing mechanism could be focused on optimizing use, rather than meeting debt-service schedules.
But of course, there is no free lunch here. By shifting revenue risk to the state, such agreements could leave the state scrambling to come up with additional sources of funds to meet its payment obligations (though with some prospect of windfalls in the out years).
Shadow tolling and availability payments are still very new ideas in the United States, despite more than a decade's use in Europe. So we don't yet know how the financial community will assess the soundness of a state's commitment to make such payments over 30 or 50 years. The Miami Port Tunnel will be an interesting case to watch.