A backlash against the privatization of toll highways was bound to occur—and we're in the midst of it today. Last year came critiques of the long-term leases of the Chicago Skyway and Indiana Toll Road. Contrary to initial criticisms that the private sector had paid too much, now the complaint was that they'd made out like bandits, and that the public sector could have generated "the same (or even greater) monetization" for these existing toll roads, in the words of critic Dennis Enright.
Now the battle has moved on to the proposed lease of the Pennsylvania Turnpike and to greenfield toll roads in Texas, such as SH 121 in Dallas. If any deal looked like an unbeatable demonstration of what the private sector could offer, it was Cintra's winning concession proposal to develop SH 121. Not only would they spend $560 million to build the toll road and operate it for 50 years. They would also provide an up-front payment of $2.1 billion, make additional annual lease payments with a net present value of $700 million, and share additional revenues if they did better than projections. And unlike the Chicago and Indiana deals, the annual cap on toll increases would be only the CPI, not the generally higher GDP per capita. What could critics possibly fault in this deal?
But Enright reappeared in April with a critique of this deal, too. Drawing on a letter from the North Texas Tollway Authority, Enright makes a comparison between the accepted Cintra proposal and a hypothetical NTTA deal to draw upon the financial strength of its whole toll road system, so as to come up with an equal $2.1 billion up-front payment. Most amazingly, he concludes that the public-sector deal could produce nearly twice as much value as Cintra's. This extraordinary claim deserves critical scrutiny, especially since it has been seized upon by opponents of concessions in Texas.
Enright's conclusion stems from several key elements of his analysis. The first is to assume that toll revenues over the 50-year period would be identical, whichever alternative is chosen. This is wrong, for two reasons.
To begin with, he adopts a traffic & revenue forecast that is unrealistically aggressive for a public toll agency. Toll agency all-debt financings rely on conservative, investment-grade forecasts. The one produced by WSA for SH 121 forecast $20.5 billion in nominal revenues, over a 50-year period. Enright uses WSA's alternative toll projection (totaling $34.7 billion), based on a more aggressive demographic forecast, which he and NTTA guess that Cintra may have used. That higher-risk forecast is appropriate for equity investors, who may not require an investment-grade rating to finance such a project. But it's unlikely to pass muster with rating agencies and tax-exempt bond buyers of an agency like NTTA, who expect investment-grade ratings. Most of the difference in value between the deals, as estimated by Enright, is due to this one factor.
The other toll-revenue problem is the assumption that a public toll agency would be able to increase tolls every year for 50 years, as authorized under a concession agreement with a private company. Political interference in toll-setting has plagued public toll agencies as long as they've been in existence. The only two cases I know of where a public agency is making regular toll increases (so far) are E-470 in Denver and the 91 Express Lanes in Orange County, California. In the latter case, the Orange County Transportation Authority has been persuaded that in order for value pricing to work, toll rates must stay at market-clearing levels, via an automatic process. As for E-470, toll rates have been increased every three years since 1995, but we have no guarantee in either case that they will be allowed to keep doing this 20, 30, or 40 years from now. I'm still very skeptical that a public agency can implement a fool-proof mechanism for ensuring the kind of 50-year revenue flow made possible by a concession agreement.
Another component of Enright's conclusion stems from his unexplained listing of the net present value of 50-year operations and maintenance costs as being 42% higher for the private firm than for NTTA. By everything we've learned about private-sector service delivery over the years, the default assumption should be that the private sector would be leaner and more efficient than the public sector.
Finally, there's the question of discount rates. In order to make a fair comparison of money flows over time, it's standard practice to use some kind of interest rate to discount future flows to present value. When a firm makes a decision about an investment, a key issue is the value of the resulting cash flows over time. From the firm's standpoint, the interest rate used reflects the level of risk associated with these future funds. An informed investor will use a different rate, depending on the nature of the investment.
This is where Enright errs. As the ultimate beneficiary, representing the public, the "investor" in this case is the Regional Transportation Council (RTC). It has a choice between two "investments": the proposal from Cintra and the hypothetical NTTA deal. Once the concession is signed, the annual lease payment from Cintra is virtually guaranteed. It has the same priority for payment as routine operating costs, and must be paid before debt service, taxes, or dividends to shareholders. But in the NTTA deal, RTC's future payments come after operating costs, debt service, and a premium that NTTA will get. A reasonable investor would be more skeptical about the value of these future payments and would assign a higher discount rate.
But Enright does just the opposite! He uses 5% for NTTA, but 6.17% for Cintra, which is his estimate of their weighted average cost of capital. This, plus his over-estimation of Cintra's O&M costs, entirely accounts for his conclusion about greater value from the public-sector deal; otherwise, his analysis shows the two deals producing equal value. But if you re-do the calculation substituting the lower (investment-grade) T&R forecast for NTTA, the private concession deal clearly produces greater value.
Besides these basic errors, this kind of comparison leaves out a crucial difference between toll agencies and concession companies: the willingness and ability to take risks. Grandiose plans to "leverage" existing toll agencies assume that conservative rating agencies and their bond-buying customers will sit quietly for massive increases in debt and adoption of very aggressive traffic forecasts. That's unlikely to happen. Concession deals are not simply the same old, same old. They are a new and important phenomenon for U.S. transportation finance.
It's unfortunate that we seem to have moved into a war between these two kinds of institution, in which one side must lose for the other to win. I've always preferred to look for win/win deals, and I still think that's possible in this arena. There are some bloated, highly politicized toll agencies out there-—no need to name names. Thriving on patronage and giving contracts to the well-connected, they waste resources and poorly serve their customers. It would be a good thing for those agencies to be privatized.
On the other hand, we have many businesslike toll agencies, with professional managers and a can-do approach that welcomes new techniques like outsourcing of service delivery--and even the use of concessions. That's what seems to be happening in Florida, where the state and metro-area toll agencies are opening the door to concession companies to take on higher-risk projects that don't fit the conservative criteria of the tax-exempt revenue bond market, but have enough upside potential to interest concessionaires. That kind of partnership would be a good way to settle the emerging toll war in Texas.