How can we ensure that government officials use their powers in the public interest?
Advocates of Benefit-Cost Analysis (BCA) lament that it is too often used simply to defend decisions that an agency has already made, rather than to inform decisions as it makes them. This is a fair criticism, but a bit naïve. If an agency did not have to defend its decision on benefit-cost grounds, why would it bother to use BCA at all? Agency heads and program managers have varied backgrounds, but typically have lots of the specialized subject-matter expertise that we hear so much about, and will often have strong opinions about what options they would like to pursue. Some of them may be predisposed to use economic analysis, but probably not very many. If they have broad authority to make decisions and if their decisions are not going to be questioned, they may simply default to some version of the “dictator” paradigm: “My own preferences are rational (i.e., transitive), so we’ll just go with those.”
But our government is one of checks and balances, rather than independent decision makers. Government is force, in the words of George Washington, and the use of force – particularly by a government against its own citizens – must be justified. Agency officials are not principals; they wield whatever power they have as agents of the people. They ought to be able to demonstrate that their discretionary official actions serve the public interest, or promote the general Welfare, or otherwise advance the common good.
As I argued previously, it is not possible to find a fully satisfactory answer to the question of what constitutes the public interest, the general welfare, or the common good – which is one reason we should limit the use of governmental force in its pursuit. But where the government is applying force, BCA can help us to distinguish those actions that appear to be justified from those that clearly are not.
Who should apply that test? It depends on the context; BCA is used for possibly millions of routine government decisions. To take one example, FEMA uses BCA software to evaluate applications for hazard mitigation grants. This might be used to decide what size culvert should be installed where a road crosses a stream, or which structures (and which infrastructures) need to be reinforced against earthquakes, or whether to build a levee around a town or move the town altogether. I helped develop that software 25 years ago, and have used it myself to complete over 200 BCAs in as many hours.
Analyses of such routine decisions, done by a competent and unbiased analyst, can usually be relied upon. But when a decision is more consequential for the agency – potentially affecting the size of its budget or the scope of its authority – some external review is necessary. BCA requires judgement calls that are easily tilted to skew the result, and no agency can be relied upon to produce an objective analysis when the result is contrary to the agency’s interest.
It is helpful to make a distinction (when possible; don’t ask me about health care!) between spending programs, like FEMA’s grant program above, and regulatory programs, like NHTSA’s CAFÉ standards for cars and trucks. In the case of spending, the use of force is generally confined to the collection of revenues that provide the ways and means to support the program. That allows us to engage in specialization. The spending agency will have the greatest expertise in its particular mission, and ( we hope) will have some enthusiasm to accomplish it successfully. If anything, it will be overenthusiastic, and will seek to spend too much unless it faces a budget constraint. The size of its budget is a good measure of the burden it imposes, and accordingly gets scrutiny from the Congress and the President.
This is how spending has been organized since 1921, when the Bureau of the Budget (then BOB, now OMB) was created. The details of spending decisions are left to the cognizant agency, subject to scrutiny by budget analysts at OMB and the staffs of congressional committees. Congress and the President participate in an annual budget process, setting agency limits, and adjusting them annually. Unable to spend as much as they would like, agencies engage in cost-effectiveness analysis that enables them to make the most of what they have. Over time and with experience, unsuccessful programs will be scaled back, while successful ones will be permitted to grow.
Well, that’s the theory, anyway. The budget process in practice is riddled with pathologies. But the notion that Congress and the President set binding limits on agency spending is unassailable. Allowing agencies to decide their own level of spending is as unthinkable as allowing employees to decide their own salaries.
Regulatory programs are entirely different. When Congress delegates to administrative agencies the authority to make binding law, the ability to use force against citizens becomes dispersed among hundreds of officials. Some have made proposals to pull this together into a kind of regulatory budget, and the debate continues on whether that is feasible or desirable, but for now agencies face no budget constraint when they use regulation to divert economic resources towards their own ends. Proposed and final rules do get reviewed at the Office of Information and Regulatory Affairs (OIRA) within OMB, where the agency’s BCA (in the form of a Regulatory Impact Analysis) gets a skeptical reading.
The fact that agency regulations are reviewed within the Executive Office of the President should be no more controversial than is the fact that spending agencies are compelled to live within a budget. Government power unchecked, whether the power to tax or the power to regulate, is intolerable. Defenders of agency “independence” and “expertise” complain that review by the President allows politics to intrude on the regulatory process. But it would be far worse to have a powerful administrative state that was not accountable to the electorate.
We should marvel at the fact that, for forty years, presidents of both parties have established BCA as the central principle for judging the soundness of regulatory decisions – incorporating the principle that regulatory agencies must act in the public interest, as best we can discern it. It is a principle that Congress should embrace and incorporate into statutory law. And, while judges cannot wade too deeply into the weeds of policy analysis, reviewing courts too will often have reason to inquire whether an agency is acting in the public interest. Benefit-Cost Analysis can help inform that inquiry.