Why Government Fails So Often: And How It Can Do Better (7 page)

†
Social Security and Medicare, our largest federal programs, are so popular that anyone who proposes to reduce their benefits risks political defeat. This is why president Barack Obama, even with the political latitude afforded by his reelection in 2012, hesitates to offer specific proposals to alter these programs, and why the proposals of bipartisan, blue-ribbon commissions from Breaux under president Bill Clinton to Simpson-Bowles under Obama have largely been ignored.

*
The concepts of public and merit goods are discussed in
chapter 3
. I discuss the effectiveness of programs to provide them throughout the book, especially in
chapter 8
.

*
Such comparisons, however, would hardly be reassuring. A comparison of our progress on seventy-five separate objectives with that of other leading democracies up to the mid-1990s found that it has been below average in roughly two-thirds of the cases and at or near the bottom in approximately half. See Derek Bok,
The Trouble with Government
(2001), table 2.

*
Clifford Winston,
Government Failure versus Market Failure: Microeconomic Policy Research and Government Performance
(2006). Winston largely eschewed studies conducted by the government, agreeing with regulatory economist Robert Hahn that such studies “can be biased, inconsistent, and technically flawed because they have not been subject to review by appropriate scholars. Hahn even suggests that some government agencies do not appear to trust the numbers produced by government assessments of their own policies” (9).

*
Winston, focusing on the academic studies that he canvassed, concedes the possibility of such a selection bias but adds that “academics, especially economists, are contentious individuals who are unlikely to shy away from the opportunity to challenge other researchers’ findings simply because they will be identified as supporting government policy.” Winston,
Government Failure versus Market Failure
, 11n3.

*
See James Q. Wilson and John J. DiIulio Jr.,
American Government: Institutions and Policies, 12th ed.
(2011), 463: “Policies once adopted tend to persist, whatever their value. (It is easier to start new programs than to end old ones.)” Congress does occasionally kill programs, but not remotely as many as it creates; Bok,
The Trouble with Government
, 91.

PART 1

The Context of Policy Making

CHAPTER 2

Success, Failure, and In Between

I
t is tempting to think that government performance, like so much else in life, is simply in the eye of the beholder. Where you stand, as Miles’s Law puts it, depends on where you sit.
*
After all, supporters of a program often perceive great success where opponents see abject failure. The truth may lie somewhere in between, as the cliché has it, but how can we know when that is so? Moreover, in between covers a vast territory, and it matters a great deal where particular programs fall within that wide spectrum.

Like investors who assess their portfolios periodically, conscientious citizens should assess their government’s performance from time to time—even though the latter is a much more difficult task, with fewer clear objectives and benchmarks. In any event, as we saw in
chapter 1
, the citizenry has already passed judgment on the federal government’s performance, and it is one of sharp disapproval. If the customer is always (or even frequently) right, then the government’s programmatic failure must be common indeed.

But if citizens are to make such judgments accurately, they will need help from policy analysts who systematically assess programs’ effectiveness. In an earlier book, I and a coauthor sought to provide such help by analyzing certain dynamics of social programs directed at unfortunate, disadvantaged people whom we called “bad draws.”
We focused attention on one criterion of policy success—how well or poorly programs
target
their scarce resources on recipients—and on two groups of recipients: “bad bets” (those who are unlikely to gain much from the program relative to others) and “bad apples” (those whose misbehavior prevents other bad draws from benefiting from the program).
¹

I have identified numerous general categories of “bad policies,” a term that reflects certain normative judgments (including poor targeting) that I shall explain later in this chapter. I list the categories here simply in order to orient readers to the many examples of success, failure, and in between that they will encounter in later chapters. These categories, which are a mix of design and performance failures, include those that are so poorly enforced as to subvert their own goals; those that entrench an inefficient or unjust status quo and retard innovation; those that permit officials to continue programs that the private sector can do as well or better and at lower cost; those that reduce or even bar competition in private goods; those that create uncertainties that discourage beneficial investment and other welfare-enhancing transactions; those that give officials too much discretion; those that give officials too little discretion; those that threaten constitutional values; those that increase moral hazard for private actors or for governmental actors
*
; those that encourage self-destructive behavior by those who can least afford it (an extreme version of the previous category); those that are costly relative to the marginal effect that the policy seeks to promote (as with bad bets); those that indirectly harm the poor by wasting resources on bad bets; those that directly harm the poor by allowing bad apples to stigmatize otherwise worthy programs; those that invite fraud and abuse; and those that redistribute wealth regressively.
²

This chapter consists of three parts. I first discuss the criteria for assessing policy effectiveness—that is, for assessing what does and does not work. I then present the most important methodology for
policy assessment: cost-benefit analysis (CBA). Finally, I provide the prescriptive context for applying CBA to actual policy decisions by elaborating fourteen normative guidelines or principles for policy makers and those who would assess their decisions.

CRITERIA OF EFFECTIVENESS

A policy, it has rightly been said, is based on a theory (often only implicit) about a causal chain linking initial conditions and future consequences.
³
A policy works to the extent that it produces outcomes that meet appropriate standards of effectiveness. This chapter develops the standards that I shall use. By the same token, a policy does not work to the extent that it falls beneath these standards. As just noted, a policy may fail for many reasons: a flawed theory or design, costs that exceed benefits, erratic enforcement, muddled implementation, unforeseen consequences,
*
or political compromises that undermine its efficacy either at its inception or as it plays out in the real world. These definitions of success and failure are tautological, of course; they beg all of the important questions about policy success and failure that I discuss in this book.

Even before addressing those questions, however, some complications must be recognized. First is
the “compared to what?” question
. Public policies should not be held up to impossible standards, and I shall not do so. Instead, I shall advance conventional, time-tested criteria which in principle are hard to dispute (though their application to specific cases, as we shall see, is often more debatable). They consider whether a policy’s benefits exceed its costs and whether it is cost-effective. This is likely to be true only if two other conditions are met: it improves on market outcomes, and is also well targeted on those who need the benefits most. These restrictive criteria are
feasible, unlike the pie-in-the-sky “Nirvana fallacy” that I criticize at the end of the chapter.

Second,
all policies succeed in some respects and fail in others.
We must not let the perfect be the enemy of the good. For our civic purposes, imperfect policies should suffice (or “satisfice,” in Herbert Simon’s decision science vocabulary
⁴
) if, but only if, it is the best that the system can reasonably do under the circumstances. We citizens have a right to demand that our policies work in this pragmatic sense. Policy effectiveness is all a matter of degree: even the best are not as good as they could be, and even the worst confer
some
benefits on
some
people, who may therefore consider it a success.
*
The reader should understand, then, that
success
and
failure
are relative terms that I shall use as a shorthand. In a democracy like ours, judgments about efficacy are ultimately rendered by those whom the policy affects as politicians, taxpayers, or otherwise.
†
As we shall see in
part 2
, and especially
chapter 8
, on implementation, what policy makers hope and promise for their programs may bear little resemblance to how well those programs actually turn out. For lots of reasons—some good, some unfortunate—the vast majority of voters do not follow these things closely. This inattention creates “political slack” that affects how policies get made, assessed, and reformed (see
chapter 5
).

Third,
policy assessment requires systematic goal clarification, fact gathering, and analysis
. Even for the well-informed, it is not simply a matter of common sense, practical wisdom, or inference drawn from a program’s durability. In their seminal book on program implementation (discussed in detail in
chapter 8
), political scientists Jeffrey Pressman and Aaron Wildavsky noted the tendency to infer,
incorrectly, that because most federal programs persist, they must have been successfully implemented and are working. As they note, “Adaptation to the environment must have been achieved; otherwise, by definition, programs would not exist. No genius is required to make programs operative if we don’t care how long they take, how much money they require, how often the objectives are altered or the means for obtaining them are changed. Indeed, the law of averages would suggest that, given sufficient new initiatives, some of them must grow and prosper in the world, though the flawed adult may bear scant resemblance to the promising child.”
⁵

Political support, while necessary to establish a policy, is insufficient to make it work. Political compromises before enactment or during implementation are often essential to the program’s very existence—and may also promote credibility and stable expectations, which are policy virtues discussed in
chapter 6
—but they can sow the seeds of policy failure. Moreover, much failure, as we shall see, is due not to these political compromises but rather to systemic reasons of the sort that I analyze in
part 2
. The political and systemic reasons why many policies don’t work are not distinct categories conceptually and are even less distinct empirically, so disentangling their effects in any particular case is very difficult. In addition, different programs mix these reasons in different proportions.

Fourth,
reasonable people will disagree about how to assess a policy
. Unanimity is not to be expected or even desired. People differ, among other things, in their predictions about what the likely effects of a policy will be; how they evaluate those effects; what relative weights they attach to them (i.e., the tradeoffs); how they discount the flow of the beneficial and costly effects over time; whether they optimize or merely “satisfice”;
⁶
and the values that they ascribe to the particular decision
processes
by which a policy is determined and implemented. In addition, as we shall see in
chapter 3
, Congress often assigns agencies several, sometimes even conflicting, missions.
Part 2
presents many examples of programs having multiple purposes that may be in tension with one another. This diversity of individual preferences and policy purposes means that any effort to
characterize, measure, and then integrate them into a single “social welfare function” (as economists call it) is bound to be arbitrary. Indeed, such a function is “impossible” in the sense that it cannot satisfy even minimal standards of consistency in most circumstances.
⁷
As discussed later in this chapter, the field of welfare economics can help us to assess consequences through the techniques of cost-benefit and cost-effectiveness analyses, but in the end, normative choices among preferences are inescapable.
*

Finally,
process norms are no substitute for policy effectiveness
. This normative diversity makes it tempting to search for more cognitively and normatively manageable standards for identifying the criteria for assessing policies, and to find them not in substantive criteria but in less controversial process norms. (As
chapter 3
notes, much judicial review of legislative and administrative actions reflects this same understandable urge to focus on process instead of substance—or at least to appear to do so.) In this spirit, public policy scholar Steven Kelman advances two standards for evaluating policy making, both of them procedural: Does the policymaking process tend to produce good public policy? Does the process itself build our dignity and our character? Renouncing any effort to define “good public policy” in substantive terms (doing so would be “excruciatingly difficult”
⁸
), Kelman focuses instead on the “public spirit” and dignitary and character-building qualities of the process. These qualities, however, are obviously conclusory and subjective. Most important, they elide what the public ultimately cares about most: what works.