Critics of public investment often suggest that, however strong the case for such outlays might be in theory, the realities of American politics make the proposals a bad idea in practice. In particular, critics argue, efforts by members of Congress to obtain "pork barrel" benefits are likely to destroy any chance that public investment programs can be effective.
Pork barrel worries are of two sorts. One is about location. Rather than being guided by a sober evaluation of where projects should best located or who can best undertake them, decisions will reflect the demands of powerful members of the congressional committees authorizing the projects or of legislators whose support was obtained by making promises about the project's location. A second fear is that once a project has gotten started and jobs are at stake, it will be difficult to shut down a white elephant investment.
The pork barrel objection is an application of what economists call the "theory of government failure." Traditional economic theory discussed "market failures," where the operation of free markets does not produce optimal economic outcomes. The suggestion was that it was an appropriate role for government to act to remedy such failures. More recently, however, economists associated with the theory of "public choice" (which applies economic models to the behavior of government) began to argue that one should not compare how markets function in practice with how government performs ideally. For, it is noted, there are systematic sources of "government failure" just as there are systematic sources of market failure. Government action might still be unjustified, these economists say, despite the presence of market failure, if the costs of government failure outweigh those of market failure.
It would be foolish not to take the pork barrel objection seriously. Certainly, members of Congress care about the location of federal projects in their districts, and they worry a great deal about anything, including canceling a government project, that will cost jobs back home. And one can certainly point to public investment projects, such as the Clinch River Breeder Reactor in the 1970s, where pork barrel considerations have been important.
Yet the pork barrel problem need not doom public investments to ineffectiveness. There are, in practice, important examples of government support for technology where pork barrel considerations do not play much of a role. The National Institutes of Health award large medical research grants, and the Defense Advanced Research Project Agency gives out large sums for high-tech research, with little pork barrel intervention. (There is some legislative earmarking for specific projects, generally driven by pork barrel considerations, but it plays a small role in these agencies' budgets.) Agencies throughout the federal government routinely award competitive contracts for providing computing products and services running into the hundreds of millions or even billions of dollars. Whatever one might say about the quality of the procurement process that backs up those decisions (and, in my own research on computer procurement, I have been quite critical of the process), it has never been suggested that those decisions are driven or even influenced by pork barrel considerations.
The pork barrel problem is worrisome, but it need not be fatal, especially if one pays attention to the design of programs and institutions involved in public investment projects. Herewith, some suggestions:
First, in statutes establishing public investment programs, Congress should create an institutional mechanism for decisions about project selection and ongoing funding that keeps future congressional involvement with the specifics to a minimum. Examples of such mechanisms include funding based on evaluation of competitive proposals by relevant agency officials using the procurement process, which is quite insulated from congressional intervention, or some version of peer review by nongovernmental officials.
Some will say that such a proposal for congressional self-denial is either utopian (Congress will never do it), or wrong (it is undemocratic), or both. But the proposal is not utopian. It is already the model used in some programs of this sort. And, as a general matter, more frequently than one would imagine Congress has indeed been willing to tie its own hand in advance out of the fear that, if it does not, it will produce what congressmen know to be bad public policy because they will find it impossible later on to resist pork barrel considerations.
The self-denial mechanism is thus analogous to the dieter who locks the door in advance to protect himself against the temptation to raid the refrigerator in the middle of the night. For example, Congress established a blue-ribbon commission to determine military base closings, whose conclusions Congress could then overrule only on an all-or-nothing basis, exactly because of worries about pork barrel objections to individual closings. Congress has adopted automatic indexing for Social Security benefits, even though indexing takes away opportunities members would otherwise gain to take credit for annual benefit increases, for fear that the constituency pressures for irresponsibly large increases would be too great.
Furthermore, keeping Congress out of selection decisions is not undemocratic, since any self-denying decision about institutional design would itself be made democratically. Indeed, it would be undemocratic to forbid Congress from creating such mechanisms.
Second, the pork barrel literature suggests that the best time to worry about issues of institutional design is at the inception of a program. Linda Cohen and Roger Noll argue in The Technology Pork Barrel that pork barrel dangers are most intense once actual projects are up and running; at that point, they have created actual rather than merely potential jobs, which would be lost if a program is shut down. Questions about benefits and costs have the greatest political weight, they maintain, during initial debates about whether to establish the program at all. Similarly, pork barrel considerations are least likely to influence issues of institutional design if they are raised and resolved before a penny is spent.
Third, chances of avoiding pork barrel problems are greatest for programs where the size of individual investments is quite small or where it is quite large, and the chances are worst for those with project sizes in between. When individual investments are small, members of Congress have little to gain from getting the program located in their district or inappropriately continuing funding. Cohen and Noll's data on congressional voting on technology projects show that when benefits to a district are low, pork barrel considerations do not sway member votes. And when the size of individual investments is large, the issues involved are likely to be of higher budgetary, media, and ideological visibility. In those cases, ideology and party affiliation matter more than pork barrel considerations. The higher the visibility of an issue, the more likely pork barrel considerations will be overwhelmed by more general policy concerns.
Finally, we need to take the admonition of the theorists of "government failure" that the costs of market and government failure be compared more seriously than they themselves often take it. Just as they criticize earlier generations of economists for automatically assuming a case for government action once a market failure has been identified, there is some tendency among public-choice theorists to assume that government action is unjustified once they have identified the presence of government failure. But if a certain project is plausible enough, it may be worthwhile to undertake, even with some pork barrel costs. As political scientist R. Douglas Arnold has written in a slightly different context, "The question of where a few thousand office workers will be located is usually secondary to the issue of exactly what they will do." If a program is good enough, it may not be the end of the world to have the work for it located in the district of the chairman of the program's authorizing committee. The costs, of course, are greater if one believes with a high degree of certainty that pork barrel politics will make it impossible to cut off expensive failures . The point is that likely pork barrel costs should be compared with the potential benefits of projects on a case-by-case basis.
It is a mistake to reject public investment projects out of hand on the ground that pork barrel politics will inevitably doom them to wastefulness. But to avoid that danger, advocates of public investment need a strategy not only for the substance of the programs they advocate but also for their institutional design.