Public Works Financing

The Pennsylvania Turnpike Numbers War

Flawed Democratic Caucus report commits serious analytical errors in its public/private comparison

"Rendell's Plan to Lease Turnpike Unwise, Study Says." That was the headline of a Philadelphia Inquirer story March 2nd, reporting on the release of a report commissioned by the Democratic Caucus of the state legislature. "For Whom the Road Tolls: Corporate Asset or Public Good" was prepared by three knowledgeable people, two finance professors at Pennsylvania State University and the co-founder of toll equipment company TransCore, now a guest lecturer at Harvard's Kennedy School of Government. Based on those credentials, I expected something more rigorous than the simplistic critiques of toll road leasing we've seen over the last two years. But I was wrong.

Although written in the style of an academic assessment, this report would not have survived an academic peer review process. It's an advocacy piece, pure and simple, that assumes away key differences between public-sector and private-sector operators, makes apples vs. oranges comparisons, and commits serious errors in its analysis. In the space allotted for this column, I can only hit the main points. But be warned: unless the toll-concessions community answers this kind of propaganda piece, its conclusions are likely to be accepted by default by legislators and reporters.

The authors' most basic flaw is to compare private-sector and public-sector operations of the Turnpike as if the main difference were in their respective cost of capital. Doing that assumes the conclusion before any data or analysis, since even if the private sector uses tax-exempt revenue bonds (PABs) for the debt portion of its financing, the higher return needed by equity providers ensures that the weighted average cost of capital (WACC) of the private sector will be higher than that of an all-debt public sector financing.

So what kinds of differences do the authors assume away? Operating and maintenance costs, for one, as if the Pennsylvania Turnpike were not the third most expensive (operating expenses per dollar of revenue) toll road in the country. Its expense ratio is 65% higher than the Illinois Tollway system and more than double that of the (larger and investor-owned) Autostrade system in Italy. Surely there is scope for significant cost reduction in the Pennsylvania Turnpike's operations. There are likely to be similar efficiencies in capital expenditures, also assumed away by the authors.

In addition, they ignore the sizeable tax benefits available to a concessionaire, saying only that "there may be a tax benefit associated with the use of capital loss carry forwards," but that quantification of this is "beyond the scope of this report." My sources suggest that on a transaction like this 50-year lease, the concessionaire might not have to pay corporate income taxes until at least year 25-that sounds like a pretty significant valuation factor to me.

Then there is outright rigging of the numbers. One gross example is using different traffic growth rates for the private-sector and public-sector cases (never explained, but clearly on display in their tables). The difference between 1%/year traffic growth used for the private sector and 2.5%/year for the public sector means the latter would have 80% more traffic by year 50, and hence much higher toll revenue. Surely that would affect the relative values.

Another real howler comes near the end, when the authors purport to present "An Apples to Apples Comparison of Present Values." They compare three cases: Corporate Lease (concession), the current plan under Act 44 (put large tolls on I-80 and transfer huge amounts of surplus revenue, along with increased Turnpike revenue, to the state DOT), and a New Jersey-type "monetization" of the Turnpike. A real comparison of the net present value of the three alternatives would involve applying an appropriate discount rate to the streams of payments to the state that each alternative would produce. In the case of the concession, that would simply be the value of the up-front payment, which is real money, check in hand, on day one. The other two would be the discounted cash flow of payment streams over a 50-year period.

But instead of doing this, for all three cases the authors use the projected payments under the Act 44 plan (which are irrelevant for the lease and the monetization) and apply, as the appropriate discount rate, the WACC they estimate for each case. Since by definition the Turnpike has the lowest WACC, it scores best. And since a debt/equity financing would have the highest WACC, it scores worst. This proves nothing, except that they can do arithmetic!

In fact, a real comparison would not discount the projected $83 billion payments under Act 44 using the Turnpike's low WACC but rather would use a very substantial discount factor, reflecting the high likelihood that the feds will disallow any transfer of surplus revenues from I-80 tolling. The only money the state can count on without the I-80 transfers is $22.5 billion from increased Turnpike tolls. Elsewhere in the report, the finance professor authors go on and on about risk factors and discounting, but fail to apply their own principles to the case at hand.

There are many other flaws in this report-misleading statements, other incorrect comparisons, and making criticisms of concessions (that high tolls will cause diversion of traffic) that apply equally to all three alternatives. Peter Samuel and I published a longer critique for the Reason Foundation, which you can read at

As for what Pennsylvania should do, my best advice is to proceed with Gov. Rendell's plan to put the Turnpike up for bid-with whatever public interest protections make sense-and see what is offered. Then compare that to the appropriately discounted proceeds of the other two alternatives. In parallel, the state DOT should go back to the drawing boards on I-80 and come up with a tolling plan that meets both the letter and the spirit of the relevant federal pilot program(s)-rebuilding the existing lanes with tolls and adding new tolled lanes. If widening is indeed needed, truck toll lanes configured for longer and heavier vehicles (such as those already legal on I-80 in Ohio and Indiana) would be the way to go.

As for the professors, Pennsylvania State University and Harvard's Kennedy School have their good names on the line. The universities should consider this "study" to be a first draft and send it out for serious academic peer review.

Robert Poole is director of transportation at Reason Foundation. An archive of Poole's work is available here and Reason's transportation research and commentary is here.

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