The prospect of $12.8 billion to invest in public infrastructure and promises that investing the payout would yield more than $1 billion every year for the life of the 75-year Pennsylvania Turnpike deal may sound like a godsend to a commonwealth filled with crumbling roads and bridges.

Aside from some grumbling about increased tolls for drivers, many regard these potential proceeds like free cash.

The truth is the money would have huge hidden costs far greater than the projected toll increases. These costs would leave Pennsylvania poorer rather than better off.

Some of the greatest expenses would come from huge unseen financial incentives used to persuade private companies to invest in highways.

Have you ever wondered why we are seeing leases running from 50 to 99 years? You should. Those long leases are not something privatizers ask for to ensure that they earn back their investment. They are required by new federal tax laws. These laws let private companies write off their investments within the first 10 or 15 years of the lease, as long as the duration of the lease exceeds the usable life of the asset. Those write-offs represent money out of taxpayers' pockets, because the private companies and their affiliates will be paying a lot less in taxes.

This isn't the only tax break created to get private companies to invest in highways. Gov. Rendell called for using a "tax-exempt, public benefit corporation under IRS code 63-20." These new 63-20 "nonprofit corporations" are nonprofit in name only. They create a shell corporation to finance privatization by issuing tax-exempt bonds, something that only governmental entities are otherwise allowed to do. Granting this privilege makes it cheaper for private companies to borrow the upfront cash, but the public pays for this private subsidy when it must make up for the missing revenues not paid to the state and federal government.

Privatization boosters count on these subsidies and the lost tax revenue remaining invisible to the public. The General Accountability Office estimates that just the federal cost from tax-exempt bonds to finance three private road projects - the Pocahontas Parkway in Virginia, the Southern Connector in South Carolina, and the monorail in Las Vegas - totaled $25 million to $35 million in 2003 alone.

These are not the only subsidies. Highway privatization deals from Colorado's E-470 and Northwest Parkway to the Indiana Toll Road include "noncompete" agreements to ensure that the private company holds a roadway monopoly within the travel corridor. E-470 and the Indiana Toll Road forbade building any "competing" roads. In the case of Indiana, that extended to any road within 10 miles. The Northwest Parkway and the proposed Pennsylvania agreement would require taxpayers to compensate the private company for lost revenue from future transportation policies that might reduce paying traffic on the company's roadway.

Pennsylvanians must also pay dearly to ensure that the contract gets enforced. Pennsylvania will need to hire expensive lawyers, accountants and consultants to ensure that private operators do not shirk from the intent of the contract. Unfortunately, the track record on privatization oversight is very poor at best, and usually nonexistent, even when it is required.

Additional upfront costs come from the difficulty of predicting a fair price for the roadway over the next 75 years. If they have done their homework, both the Pennsylvania government and the private bidders will have already spent millions in consultant fees to make projections about the cost of highway construction and maintenance, traffic levels and inflation.

With tax subsidies greatest in the first 15 years, the profit window will rapidly close, leaving the company with aging and expensive infrastructure and large debt remaining. None of the existing privatization leases has hit this point so far. All we have seen are the initial years with the big upfront payments. But when that profit window closes, there will be an incentive to let the highways crumble, sell the Turnpike at a discount, or even to walk away.

Pennsylvanians will have only unpleasant choices then - take over the deteriorated roads and fix them or build new roads. The public will find itself locked into contracts with lost autonomy and control. Pennsylvania might not even have the choice of violating the lease by building a competing road, because its mountainous topography leaves few feasible alternate routes.

All these problems are exacerbated because this deal was negotiated in secret without public input and information. Pennsylvanians now have less than a month to read and digest a 686-page contract and attempt to predict and value how its thousands of conditions might affect the commonwealth through 2084.

The public should consider these costs and uncertainties, not just the immediate upfront cash, when deciding whether privatizing the Turnpike makes sense. If a deal seems too good to be true, it probably is.