Why Government Fails So Often: And How It Can Do Better (8 page)

Contrary to Kelman’s view, I shall defend assessment criteria defined not by the policymaking
processes
, which tend to be the same whether the policy turns out to be good or bad, but by its
substantive
effectiveness, defined in terms of three normative features: its
efficiency
(its margin of benefits over costs and its cost-effectiveness); its
equity
(the fairness, by some criterion, of how those benefits and costs are distributed); and its
manageability
(the ease of implementing). As we shall see in
part 2
, these normative criteria alone will cast doubt on a distressingly large number of existing programs.

Of these three criteria of effectiveness, I shall most emphasize efficiency. All reasonable people should agree that an inefficient policy should either be improved or abandoned; this efficiency is conventionally assessed according to the familiar CBA technique discussed immediately below. Equity can often be addressed under the rubric of
target
efficiency, which finesses some difficult normative disagreements by asking not what is fair in the abstract (a judgment on which people will naturally differ) but instead whether a policy in fact maximizes
whatever
fairness goal it has adopted. I now turn to CBA and include a discussion of target efficiency. Manageability is the subject of
chapters 6
and
8
.

COST-BENEFIT ANALYSIS

Cost-benefit analysis posits that policy A is more desirable than policy B if and to the extent that the net benefits (i.e., benefits minus costs, including opportunity costs) that flow from A are larger than the net benefits that flow from B. So stated, CBA is simply rationality in the service of sound policy, a call for policies that maximize net benefits, a principle to which seemingly no sensible person could object.
It does not demand that policies be perfect, only that they be more beneficial on net than alternative policies, including the status quo.
Nevertheless, genuine technical, conceptual, and normative objections to CBA do exist, and its implementation in the real world of complex, politicized policymaking can be highly controversial.

CBA is a welfarist decision-making tool, focusing on the actual consequences of policies for human well-being.
⁹
It has gained steady support among American economists and policy experts. (It is far less accepted in Europe, where even impact assessment is viewed as relatively novel.
¹⁰
) In the 1930s, Congress required that CBA be used to assess federal water projects, and it has been widely used by
government agencies ever since. Indeed, every president from Ronald Reagan to Barack Obama has mandated its use to analyze proposed government regulations through a succession of executive orders, which established the Office of Information and Regulatory Affairs (OIRA) in the White House’s powerful Office of Management and Budget, and prescribed in some detail how OIRA is to administer the CBA process in the executive branch’s regulatory policy making.
¹¹
A large literature on CBA by academic economists and lawyers has elaborated and refined its normative justifications, valuation techniques, empirical findings, and political effects.
¹²

Despite CBA’s long-standing provenance and widespread acceptance and deployment, government’s use of it often arouses intense opposition. Some criticize the very idea of using CBA to analyze programs designed to protect values such as human life and health, biological diversity, and environmental purity that are difficult even to measure, much less price. Indeed, many argue that the very effort to quantify such values itself debases them and offends basic morality.
¹³
Others, such as policy scholar Mark Moore, show that performance measures can be developed and applied in CBA even for goals like education, policing, and public health.
¹⁴

Some criticisms of CBA are more technical or methodological in nature—that certain benefits are harder to quantify than some costs, for example. This is surely true in some cases. Thus, the reasonable accommodation requirement of the Americans with Disabilities Act of 1990 has imposed substantial, measurable costs in terms of infrastructure retrofitting, yet the benefits of inclusion to both the disabled and everyone else are difficult if not impossible to measure. Of course, the reverse may also be true. Regulations often create important, hard-to-quantify costs, such as rules that make it harder for new firms to enter. They also can harm young, unskilled, inexperienced job-seekers who are excluded from the labor market by minimum wage increases but who may not realize that those mandated increases limited their job opportunities.
*
Indeed, CBAs seldom include a rule’s
job-displacement effects, apparently on the assumption that such workers will find other jobs.
¹⁵
Another quantification challenge is how to value a life (often called the value of a statistical life, or VSL) that a particular regulation might save or sacrifice—a valuation that, like the discount rate discussed immediately below, can determine the outcome of the CBA. Again, a large literature exists on how this should be done.
¹⁶

CBA is most useful in evaluating policies in which one can plausibly hold other factors constant for purposes of the analysis than for policies that will create significant changes in factors that therefore cannot be held constant.
¹⁷
A CBA’s results may turn on significant variables whose valuations are arguably arbitrary and results-driven. Two such variables are the VSL and the social discount rate; the latter measures our willingness to forgo welfare now or in the near future welfare in exchange for more time-distant benefits, and is highly sensitive to assumptions about future economic growth and policy impacts and more significant to the extent that a policy’s costs are immediate while its benefits will come (if at all) in the future.

One much-debated approach to the discount rate is the “precautionary principle,” a highly risk-averse norm that urges a correspondingly low discount rate in calculating the present value of future benefits and costs—including those that will affect future generations—particularly those that flow from avoiding or suffering potentially catastrophic risks such as climate change.
¹⁸
An important policy question is whether the precautionary principle and CBA are compatible. Some analysts believe that they are, especially if the analysis can be tweaked to take into account diverse risk preferences, opportunity costs, loss aversion, autonomy values, and other complications. In one sense, the principle simply prescribes a very low discount rate for those policy proposals being assessed by CBA. Still, it would be wrong to underestimate the principal’s potential effect on policy assessment outcomes. In the case of climate change, for example, it could (depending on the precise discount rate used) condemn policy proposals that would increase (now or soon) the carbon burden on the environment—that is, many projects that a CBA using a higher discount rate might approve.

But the precautionary principle raises another, deeper question: Is there something about certain kinds of harms imposed in certain circumstances on certain groups of people (e.g., the unborn) that makes it socially reasonable to accord to harm avoidance more weight than a standard efficiency analysis such as CBA might give it? Although there are plausible arguments for this,
¹⁹
no clear, explicit consensus on this question has yet emerged. Even without such a consensus, particular policy assessments implicitly answer it in one way or another. In some cases, agencies like the Food and Drug Administration put an analogous thumb on the scales by building a “margin of safety” into their standards.

Many other objections to CBA arise. The data on which such assessments depend may be poor or even nonexistent, thus requiring extrapolations that may also seem arbitrary. Attributing benefits or costs to a policy often begs elusive questions such as which factors cause which effects, the baseline conditions from which gains and losses can be calculated, whether an outcome constitutes a benefit or a cost or both, which consequences are taken into account, and so forth.
²⁰
Greater consumer choice, for example, can both provide new options and paralyze decisions. More generally, the choices that underlie the numbers plugged into a CBA may leave much discretion to those conducting the analysis, so that the process generating these choices may be vulnerable to political influences.

Some oppose CBA because of its essentially utilitarian methodology, which treats all preferences the same. It gives no special weight to social values that are not expressed in people’s preference functions. It ignores the distributional effects of the policy being assessed, quantifying benefits and costs without regard to which particular groups of people receive the benefits and which ones bear the costs. It takes all preferences at face value even if they are distorted by inadequate information or other reasons. The intensity with which preferences are held—a crucial factor in coalition building, logrolling, vote trading, and other forms of political decision making—are measured in CBA by people’s
willingness
to pay. Willingness to pay, of course, is a function of
ability
to pay (i.e., wealth) and also depends
on whether one factors in other variables that affect preferences such as attitudes toward risk and felt obligations to future generations—and if so, how they are to be measured.

These criticisms and complexities of CBA reflect the inherent limitations of its methodology and data sources. All have some force; a CBA that does not take them seriously is to that extent incompetent and objectionable. Fortunately, conscientious CBA practitioners can mitigate some of these limitations. They can conduct sensitivity analyses to reveal the extent to which using different data or assumptions would affect the CBA’s outcome, and then lay out alternative scenarios based on these differences. They can specify their level of confidence in the controversial propositions that they advance. They can assure a transparent process, be explicit about the many assumptions on which the CBA rests, solicit competing assumptions and data, and respond to criticisms leveled at preliminary CBAs. They can acknowledge the fact that despite the analysts’ best efforts, at least some unanticipated consequences are inevitable, consequences that might yield a different set of conclusions. Perhaps most important, they can call attention to the normative and empirical uncertainties that no CBA can authoritatively resolve. (All of these cautions will be especially important as the Obama administration proceeds with regulations on carbon reduction, as estimates of its public health and other benefits [as well as the control costs] vary widely.
²¹
) CBA should be seen not as a decision
rule
but as a decision
tool
, one with distinctive capacities and limitations that must be compared with those of other decision tools.

To put the point differently: if factors that are less tractable to a CBA are nevertheless deemed relevant to the policy decision, then CBA cannot itself dictate the best policy, but it can at least identify these other unanalyzed factors so that decision makers can somehow take them into account. Moreover, any policy option that fails either the CBA or the cost-effectiveness analysis (its less demanding cousin, discussed above) is presumptively undesirable, leaving it up to those who accord great weight to these unanalyzed factors to overcome that presumption.

That said, even the use of these “best practices” of CBA will not forestall vehement objections to its use for decisions that affect ineffable values such as human dignity (which OIRA itself has recently sought to highlight more, as with a Centers for Disease Control and Prevention rule allowing HIV-positive people to enter the United States),
²²
the interests of future generations, or other aspects of particular assessments. After all, the greater CBA’s influence over the decision, the higher the stakes in how the CBA is conducted and the greater the possibility that any particular set of contestable assumptions in the analysis will influence the outcomes. This is altogether fitting; the contestation may well improve the quality of both the CBA that officials ultimately use and the larger decision process of which it is only a part. (Even desirable contestation may impose delay and other costs on the process, but its capacity to improve the decision may override this consideration.)

Even with all of these concessions to the legitimate concerns about CBA, some will doubtless still object in principle to using it in government decisions. But the ultimate answer to these principled objections is that CBA, conscientiously conducted with due regard for its problematic features, is probably the best that we can do to assess the efficiency merits of policies. What other decision modes are available to us in deciding among policy options? Good intentions, intuition, caprice, political power, abstract philosophical theories, or coin flips are not acceptable bases for sound policy making, although all of them (especially good intentions and political power) surely affect some decisions. As environmental economist Bjorn Lonborg has put it, CBA “is a far more effective and moral approach than basing decisions on the media’s roving gaze or the loudness of competing interest groups.”
²³
To recalcitrant objectors, then, one must respond in the spirit of Winston Churchill’s quip about democracy: CBA is the worst decision tool except for all the others that have been proposed so far.