Committee on Oversight and Government Reform, United States House of Representatives

Government Implementation of Large-Scale Projects: Government Failure, Its Sources, and Implications for the ACA Website Launch

The most extensive and contentious recent government intervention in Americans’ lives is undoubtedly the 2010 Affordable Care Act (ACA). In light of the federal government’s failure to successfully launch its website,, to implement the Act, it is useful to step back and broadly assess government’s efforts to implement and manage large projects.

In my testimony, I first discuss the conceptual justification for the government taking on large projects and what its objective should be. I then provide an overview of the available empirical evidence on the economic effects of government’s management of a range of projects and offer possible explanations for the findings. I conclude by drawing some implications that pertain to the government’s delay in launching the ACA website.


Two reasons exist to justify government implementation of large projects. The first is to correct a market failure, which could arise when a socially desirable service (that is, one whose social benefits exceed social costs) is not privately offered because it is unprofitable or requires enormous financial capital that may be unavailable in private markets. Public bus transit systems are often alleged as an example of the former and the interstate highway system is alleged as an example of the latter. Market failure also occurs when a service is undersupplied because it is a public good and susceptible to the free rider problem. National defense is a classic example of a pure public good. Innovative activity by firms may also result in free riders by creating positive spillovers to competitors.

The government can increase the nation’s welfare by financing socially desirable projects and services, including public goods, which would not be supplied by the private sector. In practice, the government can provide the service or negotiate a contract with a private firm to provide the service. In the ideal case, the government corrects a market failure and maximizes economic efficiency by setting efficient user charges, financing investments that equate marginal benefits and marginal costs, and minimizing production costs. Note that the projects and services requiring the largest investments constitute the nation’s physical infrastructure. Government has tried to spur innovation in several ways, including the establishment of a patent system and an array of subsidies for firms.

The second reason that could justify government implementation of large projects is to pursue social goals—that is, American society, like any society, seeks to solve other social problems in addition to correcting market failures and promoting economic efficiency. Those goals can be categorized broadly as attempting to reduce poverty, ensure fairness in labor markets, and provide merit goods—goods that American society believes every citizen is entitled to regardless of whether he or she can afford them, including an education, insurance against certain events that could dramatically lower the quality of life (social insurance), and protection from criminals, hostile countries and terrorists, and natural disasters.

Generally, policies to achieve those goals redistribute resources from one group of people for the benefit of another group of people, but government should nonetheless attempt to achieve those goals at minimum cost to society. The ACA arguably tries to provide a merit good and to some extent correct a market failure.


What does the empirical evidence indicate about government’s involvement in projects and services to correct market failures and achieve social goals? My 2006 Brookings book, Government Failure Versus Market Failure, indicated government’s efforts generally resulted in substantial losses in economic efficiency and missed opportunities to benefit society in a cost-efficient manner. Here I provide a brief overview and update of my findings.

The federal government, sometimes in collaboration with state and local governments, is responsible for financing and managing highways, airports, air traffic control, inland waterways, urban transit, and intercity passenger rail.  In the appendix, I present a table that summarizes the economic inefficiencies and annual welfare costs from public provision of infrastructure and urban transit that appeared in my Journal of Economic Literature, September 2013 survey of the performance of the US transportation system. The total annual cost of the economic efficiencies exceeds $100 billion. The inefficiencies are attributable to the fact that government’s provision and management of transportation services has not been guided by economic principles: prices do not reflect social marginal costs, especially a user’s contribution to congestion and delays; investments are not based on cost-benefit analysis and on accurate forecasts of costs and benefits and have therefore failed to maximize net benefits; and operating costs are significantly inflated by regulations.

The vast inefficiencies have important implications for transportation-related policies to strengthen the economy. For example, the stimulus program and ongoing calls to increase infrastructure spending must recognize that potential improvements in the nation’s productivity and employment are lessened by policy failures in the current transportation system. Similarly, the Obama administration’s vision of a high-speed passenger rail network as a transformative investment must consider costs and benefits that have traditionally been overlooked by government. Indeed, Edward Glaeser performed a series of cost-benefit calculations that were published in his 2009 New York Times column and consistently found that building such a network would not be socially desirable.  

The evidence that I report in Government Failure casts strong doubt on whether federal programs to spur innovation have supported socially beneficial programs that would have been undertaken without federal assistance. Moreover, some federal support has resulted in no accomplishments and cost taxpayers billions. The recent Solyndra fiasco harkens back to the Clinton administration’s failed effort to produce a high-gas-mileage car using a hybrid propulsion system.

Still other large-scale government projects and services have experienced serious problems including the U.S. postal system and the government’s allocation and management of public land for grazing, natural conservation, and recreational activities. The former has continued to struggle financially, with ongoing threats to discontinue Saturday service, and the latter has come under attack after the government shutdown forced national parks to close.

Finally, although I am much less familiar with empirical assessments of government services and programs to pursue social goals, such services and programs are undoubtedly not being provided at minimum social cost and are wasting a vast amount of resources.     

Explaining Government Failure

Agency limitations, regulatory constraints, and political forces combine to cause and maintain inefficient policies and to impede efficient reforms. For example, the Federal Aviation Administration (FAA) is at the heart of airport and air traffic control inefficiencies because it lacks organizational independence and is prevented to a significant extent by both the U.S. Department of Transportation and Congress from using its resources—and from encouraging airports to use theirs—more efficiently. Given that it faces opposition from two powerful branches of government, it is not surprising that the FAA finds it so difficult to reform its policies.

Government agencies do little to assess whether their vast public expenditures have been spent efficiently. Transportation officials have told the GAO (GAO-05-172) that little incentive exists for them to direct available funding to performing outcome evaluations, but they have also said that potential risks do exist from finding out that a project is not providing the intended benefits. Thus, because government measures inputs instead of outputs in many venues, transportation agencies tend to declare that a project is a success once it is operating.

Agencies are likely to have status quo bias because they may lack the technical expertise to ensure that new technologies are implemented effectively and efficiently. For example, the Federal Highway Administration has not placed a priority on using advances in information technology to improve highway travel. At the same time, FAA’s well-publicized delays in implementing new technology have tarnished its reputation to manage air traffic control effectively.

Of course, special interest politics is transparent in several areas of policy. In transportation, state and local government officials lobby for increased federal assistance for surface transportation grants and increased flexibility on how they use those funds; the American Automobile Association and the American Trucking Association have opposed efficient congestion tolls and axle-weight charges; labor unions have opposed removing Davis-Bacon regulations; and urban transit subsidies have largely been accrued by powerful interests—higher wages to labor and higher profits to suppliers of transit capital. Finally, powerful interest groups are supporting federal funding of a national high-speed rail system.

Implications for ACA Website

The potential for government failure in implementing and managing a large project should be foremost in the mind of the officials of a government agency and department when it takes responsibility for a new project. Accordingly, it is vital for those officials to take steps to anticipate and address potential failure. Based on my preceding discussion, the potential problems facing the government’s launching of the ACA website include but are not limited to:

  • Limited technical expertise and an over-reliance on contractors;
  • Little, if any, rigorous and transparent ongoing assessment because of a fear of exposing problems;
  • Status-quo bias and an inflexibility and inability to make important changes in managing a project;
  • Constraints that may affect budgeting and adoption of state-of-the-art technology.

The unfortunate result of the functionality problems and delay in launching the federal ACA website is not that the desirability of the social goal of universal coverage is necessarily reduced—the pursuit of that goal is a democratic decision that must be determined by our political system—but that the social costs of achieving this goal are already, and will continue to be, inflated. Indeed, it is my understanding that some states that produced effective ACA websites have also negotiated lower rates with insurance companies for their consumers as compared with the rates obtained by states that are using the federal website and thus did not benefit from rate negotiations. It is also possible that a state that did not produce its own website could reduce the future efficiency costs of using the federal website by arranging to pay a fee to a state that produced an effective ACA enrollment website to expand that website so people from a different state could also use its services to sign up for their insurance.

In sum, the controversy surrounding the Act should not blind policymakers to their obligation to implement the Act at minimum social costs and, if necessary, to explore alternative ways of doing so.


Inefficiencies from the Public Provision of Infrastructure and Urban Transit

 Item  Aggregate Welfare Cost ($2005)
Increasing travel delays for motorists, truckers, and shippers Cars and trucks are not charged for contributing to congestion ($45 billion excluding loss to truckers and shippers) 
Excessive damage to highway pavements Truckers are not charged efficient pavement-wear taxes for road use ($10.8 billion)
Excessive structural stress on bridges  n/a
Increasing delays for air travelers and cargo during takeoffs and landings Runway capacity is suboptimal and congestion tolls are not charged for takeoffs and landings ($16 billion); costs do include cargo
Increasing delays for air travelers in congested airspace near airports n/a
Increasing delays on waterways n/a
Highways require excessive repairs and repaving Road thickness thinner than optimal ($12.5 billion); Inferior materials are used to lay asphalt ($1 billion just for California)
Damage to cars and trucks from roads in poor condition Total damage costs to cars are estimated to be $64 billion; welfare cost n/a
Highway labor costs are inflated Federal and state regulations raise wages (welfare cost n/a)
The allocation of highway funds is inefficient Funds are not allocated to the most congested cities to minimize the cost of delays ($13.8 billion)
The cost of investments in airport runway capacity and air traffic control technology is increased by delays in project completion Regulations and mismanagement increase the costs of runway and air traffic control investments (n/a)
The allocation of funds for airports and air traffic control is inefficient Funds are not allocated to the most congested airports (ATC facilities $1.1 billion; airports n/a)
Army Corps’of Engineers waterway investments are inefficient Investments do not satisfy a cost-benefit test (n/a)
Urban transit requires excessive subsidies Fares are set below marginal cost and frequencies are excessive ($10.6 billion) “Buy American” regulations; Capital subsidies; Restrictions on releasing employees