Why Government Fails So Often: And How It Can Do Better (58 page)

The study concluded that PSRIs have contributed to the discovery of 9.3 to 21.2 percent of all drugs involved in new drug applications approved from 1990 to 2007, and that PSRIs tend to discover drugs expected to have disproportionately important clinical effects.

The Human Genome Project (HGP), which took the NIH and its associated researchers thirteen years to complete, is generally recognized as the federal government’s most important long-term investment in biomedical science. A study conducted by the Battelle Technology Partnership Practice sought to estimate the social benefits from HGP, and while such estimates are inevitably speculative to some degree, their general parameters are suggestive. The study found that the government’s investment of $5.6 billion by 2010 generated, directly and indirectly, $796 billion in economic output, personal income exceeding $244 billion, and 3.8 million job-years of employment. In the single year 2010, according to the study, the genomics-enabled industry paid back in taxes to all levels of government the
entire federal investment in the HGP. This, despite the fact that the bulk of the HGP’s likely medical, agricultural, energy, and environmental benefits still lie in the future.
¹⁰⁸

The success of NIH research seems overdetermined. It is a rarefied hothouse of scientific talent, a jewel in the federal government’s crown that attracts some of the best scientists in the world for short or long-term stints. From a purely economic point of view, powerful arguments favor public funding of basic research. Politically, powerful industries like pharmaceutical manufacturing and bioengineering support much of this research, which is an essentially free (to them), valuable input into their own businesses. Quite apart from its social benefits, NIH research funding supports prestigious institutions, creates high-end jobs, and interfaces with dynamic technology industries.

REPRISE: WHAT DO THESE SUCCESSFUL POLICIES HAVE IN COMMON?

In the end, it is hard to know for sure just why these (and other) policies succeeded where so many others have failed. No two policies are exactly alike, of course; many variables are in play. And judgments about success and failure are often contestable, as I noted in
chapters 1
and
2
. Even so, this chapter’s examples suggest some convergent reasons for their success:

Some were redistributive policies in which benefits were distributed very broadly and essentially for free. (The resources used by these programs were not actually free. All entailed opportunity costs, but these costs seemed very low at the time and were viewed as both fulfilling a moral obligation and making a sound investment for the future.) The beneficiaries of the Homestead Act and the Morrill Act received valuable land grants, education, and research services at essentially no cost to themselves. The government owned an immense patrimony of empty land (save for the relatively small, isolated number of Native Americans) that would be essentially worthless until it could be settled, cultivated, and used in ways that would enrich the nation as a whole. The GI Bill was a reward for patriotic service.

Some of these redistributive programs involve relatively simple centralized administration and correspondingly few implementation obstacles. The most important example is Social Security retirement and survivor’s benefits, which is essentially a data-management and check-writing operation. While private retirement programs are widespread, cognitive, psychological, and financial factors prevent many people from saving enough to avoid penury in old age, a condition that Social Security has forestalled.
¹⁰⁹

Some successful redistribution programs are administratively more challenging because they involve some role for the states and because of concerns about fraud and moral hazard (work incentives). SNAP and EITC are examples. For both political and program integrity reasons, these concerns are especially salient in programs that distribute valuable resources to poor people of working age. But because nourishment of dependent children, the disabled, and others is the most basic, morally compelling human need, the program’s success at guaranteeing this (as the market does not) has subordinated these concerns even as the program struggles to meet them. The outcome of the political imbroglio in 2013 in which the farm and food stamp programs are being held hostage to one another will reveal whether and on what terms the legislative bargain so essential to food stamps can be sustained.

Programs that provide public goods such as communications networks, basic research, and transportation infrastructure are the natural province of government—particularly the federal government. The NIH supports biomedical research that private firms will not adequately provide even with robust intellectual property rights.
¹¹⁰
Some national infrastructure, as with the interstate highway system and the Defense Advanced Research Projects Agency’s research, was justified as an essential national defense asset but has had far broader effects on the public. The national park system and the Smithsonian Institution are precious patrimonies (even though some of their functions are managed privately).

Some successful policies replaced earlier ones that were morally objectionable, hugely wasteful, and unpopular. The EITC and welfare
reform supplanted much-vilified New Deal and state systems of income support for the poor that voters were convinced weakened work incentives and enabled immoral conduct. The VRA ousted an oppressive, unconstitutional political regime. The Immigration Act of 1965 replaced a national origins quota system that mocked our inclusive principles, foreign policy, and economic interests. The deregulation of airlines, railroads and trucking, and energy industries displaced controls that Congress came to be persuaded were grossly inefficient.

The beneficiaries and administrators of most of the successful programs were, in an important sense, pushing against an open door. The dominant culture had already endorsed and praised the programs’ implementing actions: homesteading; establishing educational institutions; building human capital through college education; interstate mobility; the feeding of low-income children, the elderly, and the disabled; more competitive airlines; diverse immigration; and voting rights. (The NIH’s success simply required researchers to engage in more of the same old activity, and to do it better.)
To succeed, then, the programs largely needed to engage the actors’ self-interest; they did not need to create new values or transform deeply rooted behaviors
. Sometimes more forceful intervention was needed, as with the VRA, where opposition was fierce and sometimes violent. But in the other cases, the government only had to establish the entitlement or incentive, get out of the way, and let beneficiaries’ self-interests do the rest. Derek Bok makes a similar point, noting, “It is no accident that the United States enjoys its greatest successes in fields—such as the private economy, scientific research, and technological innovation—where the government need only encourage individualism and creativity by limiting restrictions on competition or by subsidizing creative people, while we lag other major democracies whenever the government undertakes more complicated tasks such as devising a healthcare system or an effective urban policy.”
¹¹¹

Some programs can only succeed by altering behavior, but some behaviors are more tractable than others. I venture that the ones that government might take on in the future are more intractable than
those it has tried to change in the past. They are the high-hanging fruit, if only because the incentives of politicians and social reforms dictated harvesting the low-hanging ones earlier. For example, it was hard enough for the 1996 welfare-to-work law to increase work effort on the part of uneducated, low-income mothers by using wage subsidies, job referrals, and child care vouchers, and for the EITC to use wage supplements to reward low-income workers. But it is much harder to alter the parenting and work practices of low-income families through programs like Family Rewards in New York City. The city’s cash payments—to induce families to visit doctors and dentists, work thirty hours per week, attend school, and improve academic performance—have had decidedly meager results.
¹¹²
Welfare-to-work and wage supplements deploy more familiar, salient, and powerful incentives. Simpler behaviors tend to be more tractable. Thus, smokers’ choices are relatively straightforward (smoke, smoke less, or quit) compared with the complex series of countless small decisions faced by the obese (what to eat, in what quantities, at what times, with what exercise, etc.). The strength of signals to people from their salient subcultures—religious communities, for example—also seems to affect their motivation and ability to alter their conduct. Peter Rossi’s Brass Law of Evaluation (see
chapter 1
) reminds us how pivotal such behavioral variables are to program success.

Policy success, then, ultimately depends on precisely what a program’s participants must do and how they must do it in order for it to work. In
Uncontrolled
, Jim Manzi reviews the experience with randomized field trials (RFTs) of proposed or adopted innovations in public policy, business, political strategy, and social science. In arguing that promising results must be replicated before being instituted as policy, Manzi elaborates on the kinds of effectiveness factors that I have just mentioned:

First, we should be very skeptical of claims of the effectiveness of new programs. Empirically, the vast majority of criminal justice, social welfare, and education programs fail replicated, independent, well-designed RFTs. Though almost any reasonable-sounding program will probably work under some conditions, most fail most of the time. The burden of proof should always be on those who claim that some new program is worth investment.

Second, within this universe of programs that are far more likely to fail than succeed, programs that attempt to improve human behavior by raising skills or consciousness are even more likely to fail than those that change incentives and environment. [Of the many ideas for pushing welfare recipients into the workforce tried in the late 1980s and early 1990s,] only adding mandatory work requirements succeeded in moving people from welfare to work humanely. And within mandatory-work programs, those that emphasized just getting a job were more effective than those that emphasized building skills. The list of both “hard” and “soft” attempts to change people to make them less likely to commit crime that do not work is also almost endless—prisoner counseling, transitional aid to prisoners, multisystemic therapy, intensive probation, juvenile boot camps, Job Corps, etc.—but the only program demonstrated to reduce crime rates in replicated RFTs across 103 documented trials is broken-windows policing, which concentrates enforcement on targeted areas and changes the environment in which criminals operate. Similarly, it is extremely difficult to find any curricular, training, or related programs that drive sustained gain in academic performance in replicated RFTs, but creating choice for students in an environment in which schools are released from collective bargaining and other constraints appears to create improvement. Therefore, we should generally seek to change incentives and environments, rather than try to change people. This is not to say that direct behavior improvement programs can never work—a program of nurse visitations to expectant and new mothers is a well-known example of a discrete program that has apparently succeeded in replicated independent RFTs—though, as with this example, those that succeed are often extensions of traditional public health measures.
¹¹³

These findings are essential to sound policy making where what is being targeted is individual behavior. It is easier to alter people’s incentives than to change their values or character. Policy environments are more tractable than the people who inhabit them. What seems to work in a pilot project run by true believers often fails when it is routinized and bureaucratized in the less rarified real world. Only field testing—as rigorous as social science can muster—can provide a
reliable basis for optimism about government-promoted behavioral change. (See
chapter 12
.)

Sometimes, external forces are simply too potent for any government program to master. For example, sixty years after
Brown v. Board of Education
was decided and fifty years after Title VI was enacted, the percentage of black children who still attend majority nonwhite schools remains roughly the same (about 75 percent).
¹¹⁴
It largely reflects old housing patterns, new locational choices, and changing demographic conditions (e.g., higher birthrates for minorities than for whites) over which the federal government has little or no control. (Indeed, its own housing programs exacerbated the original residential segregation.) We may well discover that the public’s eating habits are also largely beyond the government’s control. Sociopaths’ access to firearms,
¹¹⁵
and large-scale prevention of criminal recidivism,
¹¹⁶
seem to be other tragic examples. The persistence of these conditions is not for lack of government efforts and commitment. Instead, it is due to limits on the behavioral variables that the government can legitimately control and effectively alter.

*
For example, a leading proponent of American exceptionalism, the historian Bernard Bailyn, notes that the early colonists’ experiences “were not mainly of triumph but of confusion, failure, violence, and the loss of civility….” Bernard Bailyn,
The Barbarous Years: The Peopling of British North America: The Conflict of Civilizations, 1600–1675
(2012), xv.

*
Prevention
of AIDS, as distinguished from cure or delay, presents a different problem. Success depends on changing individual behaviors in the face of moral hazard created by the availability of a cure. See Tomas J. Philipson & Richard A. Posner,
Private Choices and Public Health: The AIDS Epidemic in an Economic Perspective
(1993).