I’m reading Peter Orzag’s statement, from 7/10/08, when he testified before the Senate’s Committee on Finance, and thought I’d collect some of the interesting bits and pieces, and offer some observations.

My testimony draws on past work done by the Congressional Budget Office (CBO) and others, and it sets the stage for more detailed analysis to identify specific economically justifiable infrastructure spending and appropriate funding mechanisms. The testimony makes the following key points:

  • Estimates from the Federal Highway Administration (FHWA) and other sources indicate that additional spending of up to tens of billions of dollars each year on transportation infrastructure projects could be justified. Some of that spending would simply maintain the current performance of existing infrastructure; other projects would improve performance to the extent that the economic benefits exceeded the costs (although some projects would have net benefits that were smaller than those that could be obtained from spending on items besides infrastructure).
  • In general, additional federal spending for nontransportation infrastructure appears more difficult to justify. In some instances, the interaction of private producers and consumers in the marketplace determines an appropriate level of spending on infrastructure. In other instances, the case for a federal role might be strong, but the case for specific additional spending either is not well documented or is difficult to justify from an economic perspective.
  • Although the rationale for some additional spending is probably strong, the economic returns on specific projects vary widely. Accordingly, even if the Congress were to increase spending, it would be important to identify which projects provided the largest potential benefit from limited budgetary resources.
  • Some of the demand for additional spending on infrastructure could be met by providing incentives to use existing infrastructure more efficiently and by devoting current budgetary resources to their highest valued uses. For example, the Department of Transportation has reported that the demand for new spending on highways could be reduced by as much as $20 billion annually if congestion pricing were implemented to encourage efficient use of existing infrastructure.
  • The question of whether projects are economically justifiable is distinct from determining who should pay for them. There is a strong economic rationale for charging beneficiaries for the costs of infrastructure. For example, it can be more efficient to impose taxes and fees on identifiable groups of users, such as drivers, than to rely on general revenues to fund an infrastructure project. Similarly, for projects whose benefits are mostly local or regional, state or local funding can be more efficient than federal funding.
  • A special-purpose entity, such as a federally chartered infrastructure bank, could provide funding for infrastructure outside of the annual appropriation process but would not be a source of “free money”: Any reduction in the federal shares of project costs (obtained by reducing grant sizes or by shifting from grants to loans or loan guarantees with smaller subsidy costs) would require greater shares to be borne by project users, state or local taxpayers, or both. …

The most recent comprehensive data, for 2004, indicate that total capital spending from all sources on transportation, utilities, and selected other public facilities — specifically, prisons, schools, and facilities related to water and other natural resources, such as dams — was more than $400 billion that year (see Table 1).1 The federal government financed about $60 billion (including federal grants to state and local governments), or roughly 15 percent of the total.2 State and local governments funded (net of the federal grants) 42 percent of the investment, and the private sector provided the balance. …

Federal spending on infrastructure is dominated by transportation, which accounted for nearly three-quarters of the roughly $60 billion total federal investment in infrastructure in 2004. Highways alone accounted for nearly half of the total.3 Capital spending by state and local governments that year was primarily for schools, highways, and water systems. Together, those categories accounted for about $135 billion in state and local government spending, which is about 80 percent of the $170 billion spent on infrastructure by state and local governments.

In contrast, private-sector investment in infrastructure is dominated by spending on energy and telecommunications, which in 2004 represented nearly 80 percent of the sector’s total infrastructure spending of about $175 billion. Private entities provide most of the nation’s electricity and telecommunications services (typically, under federal or state regulation) and account for nearly all capital spending on those utilities. …

Growing delays in air travel and surface transportation, bottlenecks in transmitting electricity, and inadequate school facilities all suggest that some targeted additional infrastructure spending could be economically justifiable. CBO’s review of the evidence suggests that tens of billions of dollars of additional infrastructure spending each year could be justified on an economic basis. The need for such spending, however, could be substantially reduced by user fees that encourage more efficient use of infrastructure. …

Although capital spending on transportation infrastructure already exceeds $100 billion annually, studies from the FHWA, the Federal Aviation Administration (FAA), and elsewhere suggest that it would cost roughly $20 billion more per year to keep transportation services at current levels. Those studies also suggest that substantially more than $20 billion in additional capital spending on transportation would be justified on economic grounds if well targeted (because such spending would generate benefits whose value exceeded its cost). …

Interestingly, Orzag’s references to justifying projects on economic grounds do not appear to take any account of the notion that every worthwhile infrastructure project has the effect of raising land values.

… according to a detailed analysis that the FHWA provided to CBO, over the next five years, investments required to maintain current levels of highway service would represent 58 percent of the total spending for all economically justifiable investments for highways, but they would provide 83 percent of the net benefits.

Table 2 on page 8 provides information about the potential for additional spending, but it provides no information about who should pay. The “benefits principle” suggests that federal taxpayers are often the least efficient source of financial support for an infrastructure investment — after the direct beneficiaries of the investment and local or state taxpayers. From the standpoint of economic efficiency, the ideal is to charge users of infrastructure according to the marginal costs of their use. For example, people who use water can be charged for the costs of acquiring, storing, treating, and distributing the water they consume.

One characteristic of many infrastructure services, however, is that some costs are not associated with anyone’s marginal use. … However, they demonstrate the willingness of users to pay for the services that are made possible by an infrastructure investment, and thus they provide an indication of that investment’s efficiency. (Indeed, the term “infrastructure demand” should arguably be reserved for desires that are supported by beneficiaries’ willingness to pay.)

Although it is generally desirable from an economic efficiency perspective, charging the beneficiaries of infrastructure investments is not always feasible, even when the benefits of such investments would exceed their costs. In some cases, the key problems are technical, such as the limitations of 20th-century methods for collecting highway tolls. In other cases, the difficulty arises because the benefits are widely distributed and preventing nonpayers from receiving the benefits is difficult or impossible, as in the case of a dam that provides flood control services. In those instances, taxpayer funding can be the most efficient solution, if the projects to be funded are chosen on the basis of benefit–cost analyses.

Those whose property is protected by that dam will have increases in their property value as a result. It should not be off-limits for the purposes of funding that dam’s construction and maintenance. Why on earth should we tax the federal taxpayer for that local bit of value?

Even with taxpayer funding, a version of the benefits principle still applies: The more closely the group being taxed matches the set of beneficiaries, the more efficient the investment decisions are likely to be. In particular, if the benefits of a project are concentrated locally or regionally, state or local governments spending their own money are likely to be in a better position to make efficient choices, weighing benefits against costs, than the federal government would be. For example, partial taxpayer support for a mass transit system could be economically efficient, to the extent that the system benefits nonriders by reducing congestion on area roads. However, decisions about the amount to invest might be less efficient if the taxes being collected come from areas that extend beyond the region served by the system.

Conversely, the case for support from federal taxpayers is strongest for investments with benefits that accrue to broad geographic areas or to the nation as a whole and are not restricted to a class of users that can be charged more directly. Infrastructure with such widespread benefits arguably includes the Interstate Highway System and wastewater treatment plants for communities whose water eventually flows into a major resource such as the Chesapeake Bay or the Gulf of Mexico. Even when federal support for a given type of infrastructure is justified in principle, implementation problems might make it undesirable in practice. If the federal government decides to channel additional infrastructure funds through state governments, some of those funds ultimately might not finance additional infrastructure; instead, federal funding might merely substitute for state and local government funding, with little or no effect on the total.

In a section entitled “Economic Returns on Public Spending on Infrastructure,” Orzag writes,

Another approach that sheds light on the appropriateness of additional spending on infrastructure reaches broadly similar conclusions. In particular, spending on infrastructure benefits the economy by reducing the cost of private business transactions; over the past 20 years, economists have attempted to measure those benefits and have obtained a wide range of estimates. The literature supports two conclusions:

  • First, public spending on infrastructure often produces positive economic returns, and second, there is significant variation — both in the average returns and in the range of returns among projects — that depends on several factors.
  • Second, the research suggests that the returns on the initial phase of a system of public investments, such as the creation of the Interstate Highway System, can be large but that the economic payoff declines as the system grows.

Federal spending on infrastructure increases the stock of publicly owned capital and, in that sense, represents an investment in the future productivity of the private sector. The economic payoff from public spending on infrastructure depends on the usefulness of the investments themselves and the extent to which the spending “crowds out” — or reduces the funding available for — investment in private capital. The early research on infrastructure spending identified substantial returns on that investment.

Think about a few large projects: the George Washington Bridge, connecting New Jersey with New York City; the Verazanno Narrows Bridge, connecting Brooklyn and Staten Island, and the Tappan Zee Bridge, connecting Rockland and Westchester County, NY, and feeding traffic to NYC. Who benefited? Those who held land made valuable by the infrastructure investment. Those on both sides of the Hudson River, for many miles around. Those who owned land at both ends of each bridge, for many miles around. And Staten Island was no longer a ferry ride from all the rest of NYC. Did we collect from them? No. We didn’t even try to record the value we created for them! What a deal! What nice people we are! We created HUGE amounts of land value, and deposited it into private pockets. And we continue to do so.

The Tappan Zee needs to be replaced. If its replacement bridge is not in place by the time that bridge is no longer safe to use, the northwestern suburbs of NYC across the Hudson will see their land values decrease, as those who live there will be caught in major traffic congestion getting to the jobs in NYC which finance their mortgage payments. Mortgage lenders will not be enthusiastic about lending there. Rents will drop off significantly.

The article, I think, gets some facts wrong in this next section, since tolls are only charged on the eastbound crossings of the Hudson River, but the discussion is instructive nonetheless:

Promote Reductions in Demand

Finally, the government could reduce the demand for additional infrastructure by implementing fees and charges that raise the cost to users of existing infrastructure.

One factor that can contribute to the high cost of infrastructure services is that users often are not asked to pay the full marginal cost of the services they use.

A classic case is the excessive crowding of a highway for which users pay no congestion charge. In economic terms, society would be better served by reducing demand for travel on such a highway during the hours when traffic is heaviest instead of investing to increase the road’s capacity to accommodate traffic. One way to reduce that inefficient demand is to impose congestion pricing — that is, to charge tolls that are higher during peak times of the day and lower during off-peak hours. Besides dampening demand for the highway during the most congested periods — some motorists would alter their travel plans and use the road when it is less crowded, find alternative routes, or switch to public transit — congestion pricing also helps to signal the places where additional investment in road capacity is warranted. FHWA has estimated that widespread use of congestion pricing would reduce by about $20 billion per year both the investment required to maintain services in their current condition and the total economically justifiable investment.

Congestion pricing is in use in the New York City area, for example, where, since March 2001, the Port Authority of New York and New Jersey has charged more for vehicles to cross the Hudson River during peak hours than during off-peak hours. The crossing’s six bridges and tunnels carry about 350,000 vehicles in each direction every day. Initially, drivers who paid with cash were charged a $6 toll, regardless of the hour of the day; drivers who used the E-ZPass electronic toll collection system paid $5 during peak hours and $4 during off-peak hours — a 20 percent discount for off-peak E-ZPass users. After the program took effect, traffic in the morning peak period declined by 7 percent from May 2000 to May 2001, and evening peak traffic declined by 4 percent (overall traffic volume remained the same).30 Six percent of trucking carriers shifted their operations to off-peak hours.31 Tolls from the Port Authority’s facilities raised $750 million in 2006, more than covering their operating and capital expenses.32 Those funds are used exclusively to build, operate, and maintain transportation facilities in the New York–New Jersey area.33 Tolls on the crossings went up March 2, 2008. The cash charge is now $8; E-ZPass rates are $8 during peak hours and $6 during off-peak hours.

Similar pricing systems have been adopted for more than half a dozen bridges, tunnels, and highways in the United States. In Orange County, California, express toll lanes built in a 10-mile section of the median strip of State Route 91 give motorists a choice between driving in toll-free lanes and driving in new lanes on which tolls are charged according to time of day. More than a dozen similar highway capacity expansions are either in operation, under construction, or in planning. On Interstate 15 in San Diego, drivers of single-occupant vehicles may pay a toll to use high-occupancy vehicle (HOV) lanes. At least a half a dozen existing HOV lanes have been converted or soon will be converted to “high-occupancy toll” (HOT) lanes.

The concept of marginal-cost pricing extends beyond congestion, however. To maximize efficiency, users would be charged for all of the incremental costs they impose on the system. For example, the incremental damage imposed by trucks on highways does not depend on a vehicle’s total weight but rather on its weight per axle.34 Because that fact is not reflected in the current taxes on truck ownership and use, there are wide disparities in the degree to which different types of trucks pay the cost of the highway damage that is associated with their use. For example, researchers have estimated that the taxes paid for a five-axle tractor–semitrailer with a gross vehicle weight of 55,000 pounds on rural interstate highways are about 20 percent more than the marginal cost of use. In contrast, the taxes paid by a vehicle with the same configuration and a gross weight of 80,000 pounds represent only one-third of the marginal costs on rural interstate highways. Marginal costs on urban interstate highways, which are more expensive to repair, or on lighter-duty roads, which incur more damage, are even higher. Instituting charges that are tied to axle weight and to the number of miles traveled by a truck could reduce the need for spending on highways by inducing motor freight carriers to reconfigure their vehicles or shippers to switch from trucks to rail. If the charges also varied by the type of road, some carriers might adjust their routes to travel on more durable roads.35

Privately run or owned toll roads

Public–Private Partnerships

Some advocates of increased spending on infrastructure suggest that greater use of public–private partnerships (PPPs) would facilitate such increases. (A PPP is an institutional arrangement in which a private entity assumes some level of risk beyond that traditionally associated with supplying its services to a government agency.) In the infrastructure arena, such partnerships appear to be most common for projects that lend themselves to private operation: roads, rail, and water supply and wastewater treatment. A private entity could control access to and charge for the use of a toll road or a drinking water system, for example, but it would be harder to charge users to recoup costs given the more diffuse benefits from a dam or flood control project.

Public–private partnerships can take a variety of forms that differ in the amount of risk assumed by the private entity. For example, private entities bidding on long-term contracts to supply services, such as maintaining public roads or operating water supply facilities, would face relatively modest risks concerning their ability to deliver services at the agreed-upon price over the length of the contract.41 In other cases, however, the private entity could have almost complete responsibility for the project and accept a variety of risks, including uncertainties about construction, the cost of financing, and the demand for the infrastructure that it provided. In some public–private partnerships for road construction, for example, the private entity could raise most or all of the funds and also would be responsible for design, construction, operation, and maintenance. That entity would recoup its investment through user fees.42

A recent report by the Government Accountability Office provides examples of PPPs for highway infrastructure in the United States, and it illustrates the use of both private management and private financing.43 Two of the four partnerships reviewed involve long-term lease concessions of existing toll roads. Chicago has entered into a 99-year lease with a private entity. That business paid the city $1.83 billion in return for the right to operate, maintain, and collect the tolls on the Chicago Skyway. Similarly, Indiana received $3.85 billion for a 75-year lease on the Indiana Toll Road. The other two cases involve plans for new toll roads. The winning bid for the first segment of the Trans-Texas Corridor (a projected 4,000-mile network of roads, railways, and utility rights-of-way) included $6 billion in capital investment for a new toll road between Dallas and San Antonio and $1.2 billion in concession payments to the state for the right to operate the facility for 50 years.44 And in Oregon, three projects have been studied under an agreement between the state and a private group to determine suitability for PPPs that would combine design services, financing, construction, and operation. Two of the three projects have been found to have insufficient toll revenue potential, but the third is moving forward to the environmental assessment phase.

PPPs have been used in many other cases to obtain private-sector financing of new toll roads, including the Dulles Greenway in Virginia and the State Route 91 and State Route 125 toll roads in California. PPPs also have been used to finance transit projects, such as the Hudson–Bergen Light Rail system in New Jersey, and freight railroad projects, including the Alameda Corridor in Los Angeles.

The potential advantages and disadvantages of PPPs include the possible reductions in investment requirements that would come with more efficient management (including cost-based pricing) and the potential increases in the costs of financing, respectively. Whether the use of private management in PPPs would help to reduce total spending on infrastructure depends on the extent to which savings from improved asset management exceed the costs of using the private services. To maximize profits, a private partner might reduce life-cycle costs through higher construction standards, more frequent maintenance, or investments in cost-saving technology. Efficiencies also could result if a private entity charged prices that were more closely aligned with costs, thereby reducing inefficient demands for services and thus perceived investment needs. However, if there is insufficient competition, public oversight could be needed to guard against the risk that the private entity might use monopoly power to raise prices excessively.

As I’ll go into in a separate post, we-the-people lose a significant source of revenue when we permit the privatization of land value, and as we contemplate significant investments in infrastructure, we ought to be focusing on how to make the effects of that investment go as far as possible, not lodge in privileged pockets, as is our tradition.