Education

In the 1980s, economists puzzled by a decline in the growth of U.S. productivity realized that American schools had taken a dramatic turn for the worse. After rising every year for fifty years, student scores on a variety of achievement tests dropped sharply in 1967. They continued to decline through 1980. The decline was so severe, John Bishop calculates, that students graduating in 1980 had learned “about 1.25 grade-level equivalents less than those who graduated in 1967.”1 Although achievement levels began to recover in 1980, the recovery has been weak and student achievement has yet to regain 1967 levels. By the turn of the century, conservative estimates of the economic growth lost due to the academic achievement decline were on the order of 3.6 percent of the 2000 gross national product.

The characteristics of the educational productivity decline challenged widely accepted educational theories of school performance. Theorists accustomed to blaming increased poverty, family instability, large class size, and insufficient spending for poor school performance could not explain why the scores of more able students declined at least as much as those of less able ones, or why measures of inferential ability and problem solving declined more than those of simpler tasks such as arithmetic computation.

Achievement fell even though the average U.S. class size shrank from twenty-seven in 1955 to fifteen in 1995.2 The decline affected students primarily after the third grade. First-graders continued to arrive at school better prepared than in any preceding generation. The decline was more pronounced at suburban schools than at inner-city ones, afflicted both private and public schools, and was larger for whites than for minorities in more-advanced grades.

The achievement decline cannot be blamed on inadequate spending. Between 1960 and 1995, annual per pupil spending in the United States rose from $2,122 to $6,434 in inflation-adjusted 1995 dollars.3 By 1999, the United States was spending an average of $7,397 per K–12 student. Spending in other industrialized countries averaged $4,850. Only Switzerland, at $8,194 per pupil, spent more than the United States.4 In industrialized countries, student scores on the Third International Mathematics and Science Studies tests are uncorrelated with spending. Though per capita U.S. spending is high and the academic achievement of its fourth-graders is above average, its eighth-graders score in the middle of the pack, and twelfth-grade achievement is consistently among the lowest of the countries studied.5

Achievement is also uncorrelated with spending within the United States. Although per pupil expenditures varied widely—from $4,413 to $13,577 per year in the one hundred largest districts in 2000–2001—no strong systematic relationship between spending and overall achievement has been demonstrated.6 Nationwide, total U.S. expenditures averaged $8,859 per public school student in 2000–2001. In comparison, U.S. home-school suppliers sell complete courses for elementary or middle-school children for less than $1,000. This price includes all books, workbooks, lesson plans, tests, and supplies. It also includes the services of a remote teacher to grade monthly subject tests; of course, it does not include the cost of parents’ time.

Given that the market price of home-school materials is about 10 percent of the average amount spent on a U.S. school student, it is not surprising that many private schools spend less than their public counterparts and turn out better-educated students. In 1990, a RAND Corporation study showed that Catholic high schools in New York City had higher graduation rates and better test scores than that city’s public high schools despite similar student bodies drawn from rough neighborhoods. The Catholic schools excelled because they elicited dramatic performance improvements from marginal students, apparently by providing more orderly environments for learning. At $3,500 per student, the Catholic schools also spent almost 50 percent less than the public schools.7

Similar patterns prevail outside the United States. When researchers in India surveyed schools for the poor in 1999, they found that actual teaching activity was occurring in only 53 percent of the public schools visited. In private schools in the same area, students paid annual fees of about U.S.$10 and classroom activity was “feverish.” The schools gave scholarships to exceptional students from families too poor to pay.8 Poor Indian parents pay extra for effective private schools because they know that their children, rather than society in general, will get most of the benefit of education and that much of this benefit will be in the form of higher lifetime wages. Unlike the wealthy, the poor cannot afford to spend years in ineffective schools.

When private schooling was the rule in Britain and America, literacy was widespread even though most people attended far fewer years of school. Economist E. G. West found that, by 1869, most people in England and Wales could read and write. In 1880, when education became compulsory in Britain, about 95 percent of fifteen-year-olds were already literate. When Massachusetts began requiring compulsory education in 1852, according to Sheldon Richman, the literacy rate was already 98 percent.9 Between 1910 and 1940, American secondary school enrollment rose from 18 to 71 percent.10 This increase is often attributed to compulsory attendance laws and prohibitions on child labor, but economists Claudia Goldin and Lawrence Katz report that secondary schools were “already largely available and free for most students” before those laws were passed.11 They attribute the enrollment increase to growing family wealth and the historically high returns to additional education.

Once public schools began using tax money to subsidize school tuition, schools financed and run by government began replacing private ones in both Britain and the United States. The number of schools fell and school governance changed radically. Because the private schools depended on student tuition, they stressed items of traditional concern to parents: challenging curricula, high academic expectations, structured environments, and efficient use of class time. As public schools crowded private ones out of the market, parents dissatisfied with one school were less likely to be able to simply enroll their child in another one better suited to their needs.

As the tax-supported schools grew larger, control shifted from parents to teachers and administrators, and the focus of school governance gradually shifted from academic achievement to staff pay and working conditions. School officials sought to avoid offending any group with political power. The result was a steady decline in academic rigor featuring, as Diane Ravitch has documented, textbooks filled with banal literature, boring writing, and inaccurate content.12

As early as 1955, some U.S. states had begun allowing teachers’ organizations to deduct dues from paychecks, bar nonunion members from teaching, and bargain collectively. The percentage of American school districts with at least 50 percent of teachers covered by collective bargaining rose from 1 percent in 1960 to 36 percent in 1992.13 The data suggest that unionization increases school budgets, and that the increase is primarily directed to increasing teacher salaries and reducing class size to reduce workload.

Most unionized school districts in the United States set teacher pay using a “single salary schedule.” Pay is determined by years of service and hours of graduate education. This equalization in pay fueled an exodus from the teaching profession between 1963 and 2000. Caroline Hoxby and Andrew Leigh found that teachers who attended colleges that required higher SAT scores for admission were the most likely to leave, presumably because they had more job opportunities outside of teaching. Overall, pay compression appears to have reduced the share of American teachers coming from colleges requiring higher average SAT scores from 5 percent to 1 percent. At the same time, the share of teachers in the lowest SAT categories rose from 16 to 36 percent.14 As Dan Goldhaber and Dan Player point out, single salary schedules waste money. Secondary school teachers with technical skills who leave teaching earn an average of $3,400 more a year in their new jobs than elementary school teachers who leave teaching for new jobs. School districts that pay all teachers the same wage either end up paying elementary school teachers too much or paying secondary school teachers too little. In the first case they waste money, and in the second they likely end up with secondary school teachers of lower quality.15

Some economists think that an increase in labor market opportunities for women means that relative teacher wages have declined and that teacher pay should be increased substantially to maintain quality.16 But wage comparisons often ignore the fact that typical teachers work fewer than 190 days a year—including days for planning, parent conferences, and professional development—compared with 235 days per year for the average full-time U.S. worker. Teaching days are also shorter. In New York and Chicago public schools, teacher workdays are less than seven hours, including a forty-five-minute duty-free lunch period.17 Teachers in public schools also enjoy lenient policies regarding unscheduled time off, low prices for generous health insurance, richer pensions, negligible travel requirements, and earlier retirements than private-sector employees. After accounting for differences in benefits and hours worked, economist Richard Vedder calculates that the average teacher earned an hourly wage of twenty-eight dollars in 2000. Computer programmers, accountants, mechanical engineers, registered nurses, and statisticians all made less.18

Current proposals for “knowledge- and skills-based” pay reforms would make public school teacher pay dependent on external pedagogical evaluations and scores on standardized tests. Like the single pay schedule, these proposals ignore labor supply and demand conditions. They also give little weight to school officials’ judgments of individual ability. State-mandated teacher certification tests are already widely in force. They appear to increase teacher wages by 3–5 percent without an observable effect on quality. It is possible that extensive reliance on time-consuming tests deters people with other job opportunities who prefer not to spend time preparing for tests.19 In private schools, both market conditions and school administrators’ perceptions of teacher quality affect pay. For example, math and science teachers at private schools typically earn 8 percent more than colleagues with similar experience.20

Teacher scores on tests of verbal ability correlate with student performance, as do the number of math and science courses taken by math and science teachers. Existing evidence suggests that variations in teacher quality account for 7–9 percent of the total variation in student achievement.21 This may not sound like a large difference. But if high-quality teachers are defined as those whose students attain higher-than-expected grades, data from individual schools in the United States suggest that five years of having a teacher at the eighty-fourth quality percentile rather than at the fiftieth would completely eliminate the performance gap between poor and nonpoor children.22

Recent efforts to identify teacher quality using “value added” methods based on rankings derived from before-and-after student achievement test scores rest on shaky statistical foundations. No one knows how to tell how much of a student’s score improvement is attributable to an individual teacher, and there is no way to know if a student who raised his score from 500 to 600 really learned less than a student who raised his score from 300 to 450.23

Many K–12 reformers seek to remedy the lack of accountability in government-funded schools by emphasizing parental choice via charter schools and voucher programs. Growing evidence suggests that competition does increase public-school quality. The introduction of charter schools in North Carolina apparently increased achievement in nearby public schools by about 1 percent, more than half of the recorded student achievement gain for 1999–2000.24 Milwaukee public schools improved after vouchers were introduced. If all schools in the United States were to improve as rapidly, estimates Caroline Hoxby, “American schools could return to their 1970–71 productivity levels in under a decade” and “school productivity might be as much as 28 percent higher than it is today.”25

Unfortunately, the American experience with voucher funding in higher education bodes ill for the long-term success of voucher funding. Grants and loans that follow students to the college of their choice, a wide choice of colleges, and providers with flexible salary schedules and the ability to hire and fire at will have not saved American higher education from recent problems with exploding costs and declining quality.

The National Association of Scholars documented the dissolution of general education requirements at the university level from 1914 to 1993.26 In a system in which bottomless taxpayer subsidies ensure that students pay only a fraction of the costs of their college education and colleges forgo substantial revenues when a student drops out due to failing grades, American colleges and universities have limited incentive to require academic excellence or to constrain spending on amenities attractive to faculty and students.

According to economists Claudia Goldin and Lawrence Katz, the modern organizational structure of U.S. higher education stems from shifts in the “structure of knowledge” that occurred in the 1800s. Government funding of colleges was originally limited to producing better-educated grade school teachers. As science became more important in business and agriculture, state funding was expanded to produce technically trained graduates of benefit to local business interests. Between 1890 and 1940, the research university blossomed, average enrollments ballooned, and the market share of private colleges declined.27 As government funding increased, its original intent was forgotten. In 1908, almost 30 percent of all students in the public universities were in engineering programs. By 1999, public engineering enrollment was no more than 5 percent.

By 2002–2003, state appropriations for higher education were about $63.6 billion. Most of the money went directly to the institutions themselves. Nearly two-fifths of the total revenue received by public colleges came from state appropriations. Private institutions received roughly half of their revenues from tuition income.28 Private for-profit degree-granting institutions got about 6 percent of their revenues from governments,29 and private not-for-profit degree-granting institutions received about 18 percent of their revenues from government through federal and state appropriations, grants, and contracts. Within this category, institutions with extensive doctoral degree programs depended on government for 28 percent of revenues, and schools offering only a baccalaureate depended on taxpayers for only 6 percent of revenues.

In return for their taxpayer funding, state college and university systems are expected to charge lower tuition. During 2002–2003, public four-year colleges charged an average list price of $4,081 per year for in-state students. Private four-year colleges charged an average of $18,273. Because students forgo state in-kind aid if they attend a private college, giving money to state colleges in return for low tuition biases students toward state institutions. In 1897, about 22 percent of higher-education students were enrolled in government-controlled institutions; in 1960 the number was 55 percent. By 1990, 68 percent of students enrolled in four-year programs were in government-supported colleges.30

As is the case in K–12 education, subsidizing institutions directly may have unintended consequences. Simulations by Ayşül Şahin of the Federal Reserve Bank of New York suggest that institutional subsidies in return for low tuition cause “an increase in the percentage of less able and less motivated college students” and that even highly able students respond by decreasing their effort levels.31

Subsidies to individual students also have unintended consequences. In the United States, mass federal subsidies for college students began in 1965, when the Higher Education Act created grants and campus programs to help poor students attend college and a loan program for middle-class students. As one would expect, federal subsidies have been systematically expanded over the last forty years in an effort to appeal to middle-class voters. More than 75 percent of federal aid to postsecondary students now underwrites loans, which appeal most strongly to middle- and upper-income students.32 In 2003, federal spending under the Higher Education Act of 1965 reached $22.7 billion. Another program, the Hope and Lifetime Earnings tax credits, totaled $4.9 billion in 2000, the third year of the program.33

As federal subsidies to students have grown, American colleges have adopted pricing schemes that maximize the revenue they receive from each student. They send a prospective student individually tailored aid packages that take into account family resources, the government aid programs available, and the student’s attractiveness to other schools. Colleges have little incentive to cut their costs and actively lobby government for more institutional funding and more student aid. Bridget Terry Long found that the new Hope and Lifetime Learning Tax Credits encouraged “students to attend more expensive colleges” and that some colleges responded by increasing tuition prices. They did not increase postsecondary enrollment.34

Aid is so generous that taxpayers are even required to subsidize child care for parents attending college. Because federal and state regulations invariably follow government funds, subsidies also increase operating costs. American colleges that admit students with federal funding are required to fulfill federal requirements on everything from reporting campus crime to meeting gender quotas in athletics.

In spite of their role in raising the cost of college, advocates of increased subsidies for higher education claim they are essential to ensuring equal access for everyone to the higher incomes generally associated with a college degree. Standard estimates of the rate of return to additional years of education range from 5 to 15 percent.35 According to a 2002 Census Bureau study, people who graduated from college earned, on average, $21,800 a year more than those who did not.36

Unfortunately, the data on average earnings from college graduation mask the fact that those earnings disproportionately accrue to people who enter college with superior cognitive skills, particularly in mathematics.37 Simulations by Hanushek, Leung, and Yilmaz suggest that need-based funding mechanisms allowing high-ability poor children to attend college may also benefit taxpayers by increasing economic efficiency.38 Subsidy programs that encourage lower-ability students to incur debt in order to attend a college that increases neither their marketable skills nor their subsequent wages harm both students and taxpayers. Daniel Boothby and Geoff Rowe report that one fifth of Canadian B.A. graduates will not earn enough extra income to recover their college costs.39

Despite lavish subsidies, only about a quarter of the U.S. population has earned a bachelor’s degree. Taxpayer subsidies for higher education generally transfer income from people with lower incomes to middle-class college students. The students capture most of the gains from attending college via higher lifetime incomes. Whether such policies are fair or desirable will likely continue to provoke intense debate.