Fault Lines: How Hidden Fractures Still Threaten the World Economy

by Raghuram G. Rajan

  Subject-matter expertise is important, but I am not persuaded that teachers need degrees in how to teach. Certainly, much of my own learning about teaching (admittedly only university students) has come from classroom experience and from mentoring by other colleagues. Requiring teachers to hold degrees in education or teaching tends to shrink the pool of candidates for teaching jobs substantially and likely deters many subject-matter experts who would otherwise become teachers. Instead, schools should create a more formal system of mentoring, with star teachers advising inexperienced teachers and sharing experiences. Furthermore, the school system should find ways to make use of those who are highly motivated and talented but are unlikely to see teaching as a long-term career. Both the young college graduate who wants to try out teaching, as in the Teach for America program, and more senior citizens, who want to give back to the community in a different role, should be welcomed.

  Small classes help the learning experience because students get more attention. This is especially important in the early years, when children need help in developing focus and discipline. Resources will be needed to reduce class size. Resources are also needed for pedagogic aids, including computers, so trade-offs have to be made based on a careful cost-benefit analysis.

  One way for the best teachers to reach more students is through technology. Technology can help teachers share experiences, lesson plans, and homework questions. I am a director of a company, Heymath, that is based in Chennai, India, and operates in three continents. The company helps teachers with the tools and templates needed to create math lessons and homework, as well as with the assessment process. It creates a community in which math teachers around the world share best practices. Heymath also offers students assistance with problems, to the extent permitted by the teacher. Technology can thus help upgrade the quality of teaching with relatively small investments.

  Schools also need a system of accountability. A national standard of student achievement, coupled with testing at regular intervals to measure performance against those standards, is key to accountability. Because schools take in students at different levels of preparation and capability, performance assessments must take into account the quality of the intake: hence performance improvements as well as absolute performance levels should be measured. We also need to find ways of publicizing school-performance assessments in a manner that is both comparable across schools and easily understood by parents. Failing schools need to be given initial support to improve, but not multiple chances to do so. The No Child Left Behind Act of 2002 goes some way toward these goals but needs to be strengthened.

  Finally, parental choice can help bring the discipline of competition to schools. School voucher programs, if properly administered, can allow students to vote with their feet and prevent failing schools from holding talented but poor students hostage. Charter schools can also help. These are quasi-public schools that have more freedom from regulation than public schools in return for greater accountability. They obtain tuition payments from school districts in proportion to the students they attract from them. Evidence suggests that such schools can lead to substantial performance improvement, especially relative to public schools that face little risk of closure.18

  The poorest-performing public schools clearly need more time to change, since the problems in the school and the community are deeper. Organizational capital and community involvement need to be built up (much as in the developing economies). If everyone thinks the school is likely to be closed, collective action is unlikely. Therefore, a period in which resources are made available, change is encouraged, and the threat of closure is held in abeyance may be necessary. But if there is no requirement for improvements in performance at the end of the period, and a permanent guarantee of a quiet uncompetitive life, it will be equally hard to elicit collective effort. As with the developing economies, a combination of initial nurturing and protection followed by competition may work well.

  The Obama administration has laid out a path for educational reform that involves a more detailed method of evaluating schools than the pass-fail system of the No Child Left Behind Act. It proposes nationwide testing standards, evaluation of teachers based on student test performance, and the entry of more charter schools, using the lever of additional cash for states and districts that adopt reforms.19 These are important and useful ideas and, if implemented, would be a major step forward in improving education. We should not, however, underestimate the resistance to common standards, transparency on performance, and student choice from failing schools and underperforming teachers, even if these constitute the minority—the attractions of a quiet life are immense and worth fighting for.

  Help for and in College

  Only 34 percent of youths from households in the bottom quintile of income distribution enroll in college, whereas 79 percent of those from households in the top quintile of family income do so.20 Moreover, only 11 percent of youngsters from the bottom quintile graduate, whereas 53 percent from the top do so. The dropout rate is thus also disproportionately high among the poor.

  Clearly, the cognitive and noncognitive skills acquired earlier are a significant source of such differences, and I discuss potential remedies above. In addition though, programs could help make it easier for disadvantaged youth to apply to, pay for, stay in, and succeed in college.

  Students from better socioeconomic backgrounds have access to significant resources to aid the college application process, including school counselors and relatives who have been to college. Students from disadvantaged backgrounds do not. Although a number of remedial government programs exist, their effectiveness is questionable.21 Recently, a number of programs have been established that link students with a disadvantaged background with undergraduate students who have recent experience with college applications.22 These deserve further study and expansion if worthwhile.

  Financial aid can also help. Studies suggest that $1,000 of subsidy increases college attendance rates by roughly 4 percentage points, improves graduation rates, and shifts students from choosing community colleges toward choosing four-year schools.23 A substantial number of college aid programs are already in place. The key to their effectiveness is to target aid to those who would not otherwise go to college, to make the aid application process simple—currently there are a multiplicity of programs, many poorly advertised and each requiring different applications, all of which pose particular difficulties for poor children without Internet access or printers—and to make the continuation of aid contingent on student performance.24

  Job Apprenticeship and Training

  Human capital can also be built through on-the-job training. Apprenticeships, especially in the private sector, can be important in inculcating work habits and behavior that can help disadvantaged youths hold a steady job. Because U.S. workers tend to change jobs a number of times in their careers, firms will have to be given incentives to offer these apprenticeships to compensate for the facts that they are unlikely to keep an apprentice in the long run and that such apprentices will require a fair amount of help and supervision.

  Determining What Works

  Many good ideas have been, and are being, proposed for improving human capital for the disadvantaged. However, ideas that make sense on paper often do not seem to work in practice. Although we should try ideas that evidence and common sense suggest stand a good chance of success, ideally, programs should also be subject to periodic evaluation. These evaluations will help identify ineffective programs, as well as modifications that could make a program work better.

  As the quality of the human capital of the disadvantaged improves, wage inequality should diminish. There is no certainty, however, about whether any interventions will be implemented, whether they will work, and, if they do, how much time they will take. There could well be political pressure from those who have fallen behind to redistribute, either directly or by the extension of more easy credit. Countervailing pressure will come from those who have to pay the highe
r taxes or who do not believe in income redistribution. Political strife could well continue at an elevated level for a while. We must hope, however, that as the disadvantaged see the ladder of programs that will help their children climb to a better future, the prospective redistribution of opportunities will head off potential conflict and the associated costs to growth.

  Security and the Safety Net

  As I argue in chapter 4 and chapter 5, the absence of a sufficient safety net to support workers who lose jobs in a prolonged downturn in the United States increases public anxiety and tends to generate a disproportionate monetary and fiscal response. The remedy is to improve workers’ resilience to economic adversity. This issue is not unrelated to that of inequality. Low incomes give people little margin of safety or the ability to save enough to tide themselves over a period of unemployment.

  In searching for remedies, we should recognize that the nature of firms and the employment relationship in the United States, which emphasize flexibility and mobility, may not sit well with the kind of long-duration safety nets available in continental Europe. In what follows, I examine the broad changes that could make the U.S. worker more resilient in downturns while preserving flexibility in the system.

  Contingent but Predetermined Unemployment Insurance

  The current unemployment safety net (beyond the minimum in place) is a contingent one whereby politicians respond to prolonged, massive unemployment by temporarily extending unemployment benefits. Workers experience tremendous uncertainty over whether benefits will be extended and the criteria for receiving them. Moreover, politicians exploit the public demand for action to push through all manner of pet projects without much scrutiny under the guise of emergency legislation.

  If U.S. recoveries have indeed changed in nature and become more jobless, it is worth debating whether unemployment benefits need to be extended on a more permanent basis. Longer-duration unemployment benefits not only entail costs but also make some of the unemployed less anxious to find jobs. A less responsive labor market may alter the nature of the employment relationship in the United States as well as the ease with which firms can take up new opportunities. The United States may want to give up some corporate flexibility for greater worker security, but that decision, because of its long-term ramifications, can be more fruitfully addressed once we have a better analysis of whether recoveries have indeed changed, and, if so, why.

  There are, however, two reforms of which the net benefits are clearer. First, the United States would benefit from predetermined, contingent extensions of unemployment insurance. In other words, instead of extending unemployment benefits on an ad hoc basis according to the politics of the moment, the decision could be tied to a formula. The formula could be some simple function of the extent of overall job losses (as already offered by some states), the proportion of jobs created to jobs lost, and the time elapsed since the recession began. More complicated formulas could take into account the sectors where jobs are lost (are these industries where output tends to be cyclical, or do the job losses indicate a more permanent transformation?), as well as the nature of jobs created (are these full-time, permanent jobs or part-time and temporary ones?). The virtue of keeping it simple is that extensions could be easier for workers to forecast and plan for, and such foreknowledge is central to reducing anxiety.

  The second area requiring change is health care. Much of the anxiety surrounding unemployment has to do with employees’ concerns about losing health insurance for themselves and their families. As I write, Congress has just passed a bill to create near-universal health care—near-universal because illegal immigrants will still be left uncovered even when the bill’s provisions are fully implemented. The passage of the bill is just the first step in a long and arduous process to achieve effective universal health insurance. Indeed, with many of the provisions of the bill due to be implemented only years from now, with attempts by various state legislatures to opt out of the provisions of the bill, and with threats by conservatives to repeal the bill when they gain power, there is some small uncertainty still about whether even the main legislative hurdles have been overcome. Even if the legislation survives challenges, effective reform will have to make sure everyone is covered even while slowing or even reversing the rate of growth of health care costs. However, the political calculus, as reflected in this bill, is to emphasize the clear and immediate benefits of expanded coverage, while providing less detail about costs and controls. There are many good ideas in the bill (and some bad ones), but only time will tell which ideas gain momentum. The debate about universal health care and health care costs is far from settled, and it will be with us for many years to come; hence a closer look at the key issues is useful.

  Universal Health Care

  There is little appetite in the United States for a radical overhaul of the largely private health care system. The key to universal health care, then, is to deal with the adverse-selection problem. If an insurance plan attracts a disproportionate number of those with preexisting health problems, which make them higher insurance risks, while attracting too few of the young and healthy whose premiums are necessary to subsidize the less healthy, the economics supporting insurance breaks down, and it becomes uneconomic.

  Given that a number of the young and healthy prefer to stay uninsured—the young have a strong and misplaced sense of their own immortality—the key to successful universal insurance, in which no private plan can refuse those with preexisting conditions and premiums are kept at a reasonable level, is to ensure that the young and healthy pay in. Incentives can include a mix of carrots—such as rebates for not making claims—and sticks—penalties for not joining. These have to be high enough to deter healthy individuals from paying the penalty and joining the system only when the need arises. Moreover, plans that attract a disproportionate number of high-risk individuals need to be compensated with transfers from those that don’t.

  Universal health insurance will also require subsidies for the poor. This requirement clearly creates some tension with the American aversion to unconditional redistribution that I discuss in chapter 4. Yet it is implausible that many Americans truly believe that their fellow citizens who have bad luck, or earn little, should pay with their health or the health of their children. Moreover, many of the uninsured sick eventually do receive treatment, but without any of the benefits of preventive care, thus exacerbating both health problems and treatment costs. Finally, there is a broader benefit to universal insurance. The reduction in anxiety among the population in downturns, and the associated reduction in the need for expansionary macroeconomic policies, will make the United States better able to calibrate fiscal and monetary policy to its actual needs, while serving notice on policy makers in the rest of the world that they need to become more expansionary.

  The major problem in expanding coverage, once the adverse-selection problem is dealt with, is that U.S. health care does not seem cost-effective. The United States spent 15 percent of GDP on health care in 2006, compared to 11 percent in France and Germany, 10 percent in Canada, and 8 percent in the United Kingdom and Japan.25 On a per capita basis, the United States spent $6,347, while Japan spent $2,474. The difference has been growing: in 1970 the United States spent 40 percent more on health care per capita than peer countries; in 2004 this difference had widened to 90 percent. But health outcomes have not improved commensurately: every country in the United States’ peer group had a higher increase in life expectancy at age 65 over the period 1970–2004, except Canada, which experienced the same 3.7-year increase.26

  Three factors seem important in driving up costs.27 First, the United States has high prices for inputs: hip replacements in the United States, for example, cost twice as much as in Canada. Part of this difference is explained by higher compensation for doctors in the United States.28 Second, because doctors and hospitals get paid for services provided rather than for outcomes, and because insurance picks up the bill, the system tends to promote overutilization of hea
lth care. Finally, the system tends to adopt innovations even when the evidence of effectiveness is weak. For instance, nuclear particle accelerators, costing more than $100 million, are very effective in treating rare brain and neck tumors; but they are also used to treat more common prostate cancers, with little additional benefit.29 A good bill would encourage more competition, a move to paying for outcomes rather than inputs (as well as setting the per-visit deductible cost to each user at a level that encourages sound judgment about whether to visit the doctor), and a greater focus on using only proven treatments that offer significant improvements over cheaper alternatives.

  Three other factors contribute to the high cost of health care in the United States but are not central. The first is administrative costs. These are notoriously difficult to estimate precisely, but one estimate indicates that the United States spent $465 per capita in 2006, while Japan spent only $52.30 Clearly, given that the overall difference in health care costs between the United States and Japan is about ten times the difference in administrative costs, they are not the sole explanation. However, measures to improve cost efficiency are important. The information technology revolution could finally be brought to health care by, for example, standardizing electronic patient records so that they can be transferred easily among different providers.

  The second is operational efficiency. Hospitals in the United States could learn more from each other, as well as from hospitals elsewhere, including India, where costs have been brought down by bringing mass-production techniques perfected in manufacturing to health care. Indian hospitals have found that error rates are reduced when their doctors specialize and perform many procedures of a similar kind. The time for operations is also cut down, with no loss of safety. A focus on eliminating unnecessary frills and on utilizing expensive resources like doctor time most effectively also helps: even though good surgeons in India earn about as much as surgeons in the United States, the cost of operations is often an order of magnitude lower. Regulations that force hospitals in the United States to be “full-service” hospitals rather than permitting specialization tend to drive up costs. Greater competition between hospitals could also bring down costs; an easy way of encouraging cross-border competition is to authorize Medicare and Medicaid reimbursements for procedures performed by authorized hospitals in other countries, like Mexico and Thailand.