Why Government Fails So Often: And How It Can Do Better (30 page)

Occasionally, policy changes are convulsive enough to violate the Constitution. In the most dramatic example of this, the Affordable Care Act amended the Medicaid program, a huge share of the states’ budgets, to sharply increase the costs to the states. A Supreme Court majority struck down provisions that would have eliminated all
Medicaid funding for any state declining to participate in the now costlier program, calling this threat “a gun to the head.”
⁷²
The main point here is not that the Court was correct to invalidate those provisions (I have serious doubts, and four justices strenuously dissented) but that the government often reconfigures (or abandons) programs that it previously urged states and private interests to rely on and invest in. The need to avoid both the Scylla of abrupt policy changes that violate government-created expectations and the Charybdis of maintaining the status quo long after new conditions demand its reform is precisely the dilemma noted above by Rodrik and Zeckhauser.

Economist Amihai Glazer and political scientist Lawrence Rothenberg have analyzed the political economy of government credibility. Where government regulates a concentrated industry, the authors find, the credibility of its policies is especially low because the industry is better able to coordinate efforts to delay or defeat those policies—as when the auto industry was able to delay passive restraints and fuel economy standards for many years, thereby discouraging private investments in the necessary supporting technologies.
⁷³
New evidence shows the same to be true of investments in renewable energy generation.
⁷⁴
Glazer and Rothenberg make two other points about credibility. First, government cannot credibly ration services whose denial would be visibly life-threatening, as with health care.
⁷⁵
Second, weak government institutions can actually increase credibility by making policy change more unlikely—that is, more rigid—and this may sometimes be more important than adapting to changed conditions.
⁷⁶
Rodrik and Zeckhauser suggest how government might enhance its credibility—giving “hostages” and creating countervailing constituencies, for example—but such techniques involve trade-offs. Each would reduce a policy’s effectiveness, increase its cost, or limit the government’s responsiveness to new conditions (a special kind of cost).

One type of institution through which Congress seeks to increase government credibility is the statutory trust fund. The trust fund earmarks certain taxes for the fund’s use, and it can only spend those revenues for specified fund purposes. Congress has used this
device in more than 150 programs; the most prominent are Social Security, Medicare, and highway and airport construction. Other statutory precommitment arrangements include entitlement programs, independent agency status, and indexation formulas. Eric Patashnik’s authoritative study of the trust fund device finds its propensity to increase government promise-keeping depends on a number of design factors, especially the extent to which the fund’s taxpayers and beneficiaries are the same people (“reciprocity”) and depend on the promises being fulfilled (“reliance”). Social Security and Medicare score highest on both reciprocity and reliance, and their commitments have (so far) been kept. For most trust funds, however, Congress has changed the rules so as to avoid having to keep the original policy promises.
⁷⁷
To avoid a fiscal catastrophe, Congress must certainly change the rules for Social Security and Medicare as well—the only questions are when and how—and much of the public has lost confidence in the programs’ ability to discharge their existing commitments.

For our present purposes, two key points emerge. First, the credibility of a trust-funded policy depends on its underlying political structure. Second, the device may increase a policy’s credibility at its inception, but that enhanced credibility either is unwarranted (for most trust funds) or tends to erode over time. Even the most reliable ones, Social Security and Medicare, are undermined by policy makers’ tendency—grounded in their electoral incentives—to overpromise benefits and undertax toward paying for them.

Although the weak credibility of public policies inheres in the government’s duty to respond to changed conditions, the problem is greatly exacerbated by a number of other factors that cause government to actually magnify uncertainty. Two of them are discussed at length in other parts of this book: the nature of its institutions (
chapter 3
), and the political culture in which those institutions operate (
chapter 6
), including the growing politicization and protracted delays in Senate confirmation of key executive and independent agency policy makers (
chapter 10
). Still other factors include the long lag times between identifying a problem and implementing a policy
response, and the political conflict that continues long after a policy’s enactment. Examples of such government-exacerbated uncertainty abound, and some of them have exceedingly far-reaching negative effects. By almost all accounts, the uncertainty caused by the “fiscal cliff” that loomed in 2012 reduced private-sector spending and hiring at a time when both were desperately needed, with baleful economic and social results.
⁷⁸
Much the same is true of the “regulatory cliff” that pushed off final action on a number of major costly regulatory actions until after the 2012 elections.
⁷⁹

The Affordable Care Act of 2010 is an even more revealing example of policy-driven uncertainty because the law has already been on the books for over three years. The act’s implementation depends upon innumerable public and private decisions—to invest, to legislate, to establish exchanges, to coordinate databases, to insure or exclude, to contract, to organize, and countless other actions—and the effectiveness of its many different policies turn on when, how, and how well those decisions are made and integrated. Yet for more than three years, its very existence, not to mention its authoritative meaning and timely implementation, rested in the opaque, unpredictable, and formally unaccountable hands of the U.S. Supreme Court—of one or two “swing” justices, actually. This has created a mind-boggling level of uncertainty throughout the relevant industries and patient populations.
⁸⁰
And even when the Court finally rendered its decision upholding most of the act,
⁸¹
it did so in a way that created new uncertainty about both the federal program and the continuing status of fifty state Medicaid programs for the poor. At the same time, the Republicans (and a few Democrats like Senator Joe Manchin) moved vigorously to repeal, delay, or seriously amend the Act at the first political opportunity. Their arguments were aided by the computer system’s egregious failures in the first months of operations, failures that the Obama administration had struggled mightily to avoid.
⁸²

This uncertainty will not be resolved for years to come, yet a vast number of public and private insurers, providers, consumers, consultants, and other actors must quickly undertake make-it-or-break-it
decisions based on speculative possibilities about which rules numerous officials in different federal departments and state agencies will negotiate among themselves and then issue.
⁸³
In September 2013, only one month before the act’s key provisions on insurance exchanges went into effect, even the major labor unions that had supported it expressed deep concerns about how the act could harm their members.
⁸⁴
Here are just a few of the myriad questions that as of July 2013 remained unanswered (which also meant that the necessary computer systems could not be designed and field-tested):

To comply with the law, companies needed rules for reporting data to the government each month, including the names and Social Security numbers of anyone who worked full-time for at least one month during the year, plus information on the insurance offered and its price. They have to calculate and report, based on individual worker incomes, whether the premiums offered are affordable. Among the questions the administration hadn’t answered in time: should companies be required to report month-by-month details about who they employed? Should employers who already offer insurance be subject to the same rules? What happens if workers say they weren’t offered adequate insurance but employers say they were?
⁸⁵

As of late November, even the President conceded that the initial rollout was terribly flawed. Other uncertainties abounded quite apart from the Republican’s quixotic threat to repeal or fundamentally change the act. Most states were operating their own insurance exchanges. Many young, low-risk people on whom the law’s viability depends might decide not to get insured at all or instead pay the low statutory penalty (the final rules were issued only a month before the insurance exchanges went into effect).
⁸⁶
How strictly the government would enforce that penalty was unclear, as were the premiums that insurers would charge. Much, much more was still up in the air.
⁸⁷
When the administration suddenly suspended the employer penalty provision for a year because it could not yet predict how employers, insurers, and workers would be affected and respond,
⁸⁸
it not only cost an extra $12 billion, according to the CBO,
⁸⁹
but it created yet more confusion: whether the individual mandate might also be
delayed, how agencies would get the information they needed (and apparently still lacked) to implement the many interlocking provisions, and so forth. Tasting blood, the program’s opponents began to press for additional suspensions.
⁹⁰
The program’s credibility was further damaged when, only six weeks before rollout, another important delay emerged—this relating to the rules limiting patients’ out-of-pocket costs.
⁹¹

If a landmark social policy (surely the most important since the New Deal, other than Social Security and Medicare) must contend with this level of uncertainty—if government promises about the act’s content therefore lack credibility in the minds of insurers, providers, patients, state legislatures, and countless other stakeholders whose own commitments are essential for success—what does this portend for the act’s ultimate, overall effectiveness? (I say “overall” in recognition that
some
of its popular goals—especially increasing coverage of those now barred by preexisting conditions or by high individual policy costs—will likely be realized, albeit at a greater-than-predicted cost.)

The question, alas, answers itself.

MISMANAGEMENT

Sound management is essential to policy effectiveness, and we must presume that much more of it exists than a scandal-hungry media tends to report. Indeed, an instance of good management can become newsworthy, a man-bites-dog story.
⁹²
Nevertheless, the problem is not media bias but endemic federal mismanagement that can have terrible consequences. Consider three particularly appalling examples. First, as many as half of discarded kidneys could be transplanted if the federal Organ Procurement and Transplantation Network were not hobbled by an outdated computer matching program, red tape, overreliance on inconclusive tests, and even federal age discrimination laws.
⁹³
Second, the Drug Enforcement Agency had to pay $4 million to a student who it mistakenly jailed for four days and
who almost died as a result.
⁹⁴
Third, the popular policy of providing generous benefits to disabled and ill veterans, our most honored citizens, is mocked by the Veterans Administration’s long claims-processing delays, which even predated the 9/11 attacks. The VA’s $140 billion budget has more than doubled during the last decade in real terms, and it has added thousands of claims processers. Yet as Congress keeps authorizing new benefits and makes eligibility easier, the backlog (now 900,000 claims) grows steadily worse due to the agency’s continued reliance on paper records, its perversely designed production quotas that encourage employees to reach for the thin folders first, the numerous refiled and appealed claims after denials, and its lax definition of disability to include common age-related conditions.
⁹⁵
The
New York Times
reported that even a routine pension claim, undisputed by the VA, took nearly two years to process, and only after a congressman’s intervention. The VA’s employees publicly, and at some personal risk, criticize their own agency as dysfunctional; the agency has decided to address the backlog through “provisional” rulings to old cases, which will likely assure payments with little or no scrutiny.
⁹⁶

Another form of managerial failure is program fragmentation and overlap—a problem largely originating with Congress. Bureaucracy expert Paul Light notes the duplication and overlap among programs with nearly identical missions but separate overhead, including the 53 designed to spur entrepreneurship, the 82 to improve teacher quality, the 160 to support housing, and the 209 to strengthen science, technology, engineering, and mathematics education.
⁹⁷
A 2011 GAO report identified forty-seven separate federally funded job training programs administered across nine agencies in 2009. Almost all of these programs overlapped with at least one other program providing at least one similar service to a similar population; differences sometimes existed in eligibility, objectives, and service delivery.
⁹⁸
Indeed, Congress apparently added two more such programs
after
this critical GAO report.
⁹⁹
In 2008, Congress created a new US Department of Agriculture (USDA) office to inspect imported catfish, which is a
low-risk food, duplicating a long-standing FDA seafood inspection program. By 2013, the USDA had spent $20 million to establish it and $14 million a year to run it, yet it had not inspected a single catfish.
¹⁰⁰