Robert Frank makes an interesting case for investing in infrastructure now. If you are of the Keynesian persuasion, even the most worthless infrastructure investment makes sense these days -- anything to put paychecks in people's pockets, even if financed purely by debt.
Suppose you don't buy the Keynesian argument. There's still a case to be made that this is a good time to engage in productive investment -- not digging holes and filling them back in again, but actually building or repairing stuff we'll be happy to have later. Interest rates could not be much lower. You could probably cut a pretty good deal on construction costs. It's not crazy to think that any project worth contemplating over the next 10-20 years shoud be fast-tracked and started now.
But what projects are worth contemplating? Currently, we don't really have a rational mechanism for picking out projects that would be good investments -- those with a sound rate of return. We end up building bridges to nowhere, trains to nowhere, and the like. The appropriations process gives us things like the Bud Shuster highway, while relegating citizens in less-favored districts to infrastructure that is outdated and congested at best, fatal at worst.
Is there a more rational way to channel our investment dollars, toward projects that pay for themselves and away from boondoggles? Can we take politics out of the equation entirely? Maybe.
Here's the basic formula. The nation simply needs to do two things: 1) establish a list of prioritized projects, 2) prepare to pay tolls for those projects that are completed. Tolls are bad, I know, but the alternative is to pay with tax dollars, which isn't any better and from the economist's perspective in fact worse. Would you rather have the next bridge to an Alaskan island financed with your tax dollars or with toll revenue from those who actually use the bridge? Congestion tolls can actually help reduce traffic problems and car exhaust emissions.
We then rely on the private sector to establish which prioritized projects actually stand to pay for themselves. For each prioritized project, the government offers a contract to construct or repair, then to collect and keep toll revenue for a period of time. The contract can specify quality standards and even maximum tolls to be levied. Prospective contractors then bid for the contracts, not in terms of dollars but duration. The cheaper the project, or the higher the maximum toll, the longer the contractor will bid to keep the road. The bidder who offers to turn the project back over the fastest wins the contract.
This is not an exotic auction format. These are build-operate-transfer (BOT) contracts, already used in several states and foreign countries. Contractors may be nervous about bidding for them: after all, if toll revenues come in below projections, they may lose a lot of money on the deal.
This leads to another issue. Firms bidding on a BOT contract will generally take out their own loans to do the work, which misses the whole point of taking advantage of low government borrowing costs. The government could lend money to the contractor itself, issuing bonds to pay for it, but then may end up holding the bag if the contractor can't make enough money to pay off the loan. One could also consider government loan guarantees instead of outright financing, but that carries the exact same risk. The best you could do would be to mitigate the risk. If the contract is detailed enough, it would be relatively straightforward to determine which bidders have the most optimistic attitude towards future traffic -- or, more cynically, toward the likelihood of a future bailout.
So, alas, there is in fact no such thing as a free lunch. But the BOT mechanism would at least put the focus where it belongs in infrastructure investments: on rates of return.
