Education reform as economic reform.

• Jurisdiction: United States
• Date: 22 March 2005
• Author: Osborne, Evan
• Published date: 22 March 2005
• Author: Osborne, Evan

In 1642 the Massachusetts Bay Colony enacted an ordinance pronouncing, for the first time on these shores, that public education was a fundamental mission of the state. The Law of 1642 expressed concern over lack of knowledge of English and of the "Capital Lawes." By 1647 towns in Massachusetts of at least 50 people were required to hire a school master, because the "Old Deluder Satan" was prone to tempting children into ignorance of the scriptures. Education has thus since the earliest days of the country been seen as a public function in the United States. In the founding era, Jefferson was the most famous advocate of broad public schooling. But it was not until the era of industrialization and mass immigration in the latter portion of the 19th century that public schools, from grammar school through the universities, became an imperative in most states. The need to acculturate immigrants and to prepare citizens to participate in an increasingly sophisticated economy was a primary driving force in this movement.

But at least since the publication in 1983 of "A Nation at Risk," American public schools have been as notorious for their flaws as they once were celebrated for their necessity. What are seen as the myriad problems of public schools have drawn much comment, often including highly detailed solutions to what are claimed to be highly specific yet complete causes of these problems. The maze of competing diagnoses and proposals makes education reform a difficult problem to study. But reforming education shares some features with the problem of reforming economic policy, a widely studied problem. Economic reform seeks to improve economic growth and hence the opportunity available to residents of the reforming country, and educational reform similarly seeks to better provide children with the skills that open up more choices to them. This article applies some of the findings of the literature on economic reform to draw lessons about repairing American public education.

Taking the Measure of American Education

U.S. education may in theory be measured against some hypothetical ideal, but constraints afflicting collective education apply in all countries, not just in the United States. The performance of U.S. students relative to that of students in other countries is revealing, in that a poor U.S. performance would indicate problems peculiar to this country. The National Center for Education Statistics has summarized results of international performance on standardized tests in science and mathematics (NCES 1997). The standardized mean scores for each country are reported in Table 1.

The results are striking in one sense. American students begin above average, and decline relatively throughout. Every day in an American school seems to cause the average American student to fall further and further behind his foreign counterparts. This is a rather striking finding. It suggests that the problems in American schools are not related to the higher levels of poverty and inequality that by some measures prevail in the United States and that in theory bring with them many substantial handicaps to school performance. Nor does poor American performance appear to be related to differential family structure, religious objections to modern science, or any of the other usual suspects often trotted out as explanations. There is no reason to think that these problems begin to take hold late in a student's educational life.

Rather, if these were significant causes of poor American performance the relative gap should be similar throughout the school life. Note that black American achievement measures relative to those of white Americans are relatively constant as students move through the school system. Data from the National Association of Educational Progress show that in mathematics, blacks' scores are 87.1, 86.3 and 88.9 percent of whites' scores in 4th, 8th, and 12th grade, respectively. In science, the relative scores are 78.0, 75.6, and 79.7 percent. Thus, it appears that in international comparisons American students do not start out doing poorly, but they end up that way. The problems appear to be in the schools rather than in the broader society.

Spending across countries does not appear to be a likely candidate either, because like the aforementioned other factors, its effects should be as salient at each stage of the schooling process. If there is a monotonic relationship between relative spending and relative performance, then if the United States spends 20 percent less than Japan (for example) the gaps in relative U.S. performance should show up at the earliest levels. The United States certainly should not be in the higher strata in the early years. As a cheek, Table 2 presents multinational data on both total public and private spending on education as a percentage of GDP and per student in 2000 (OECD 2003) and the total number of school hours per year. (2) These are measures of inputs into education. The United States is not a particularly low-spending nation, ranking 14th out of 35 nations in spending as a percentage of GDP. Considering that the United States is among the wealthiest OECD nations, per-student expenditure is actually quite large, as can be seen in the second column, where the United States ranks second only to Switzerland out of 25 nations. The time input is also apparently not responsible. While the United States does have a short school year measured in terms of days (38th out of 42), it ranks close to the top in total number of hours per year (4th out of 27). While the inconsistency with respect to these two time measures is somewhat surprising, it is identical to the finding in Matheson et al. (1996).

In short, the United States spends a substantial amount on each of its students, and while the students spend a reasonable amount of time in school, the more years they are there the worse they do relative to other countries. Given the dominance of public schools in American education, it is worth exploring their performance in particular. The TIMSS data do not break down performance between publicly and privately schooled students, but comparisons can be made within the United States, using other data available from the NAEP and presented in Table 3. The total gap, in the top row, between public schools (not including Defense Department and Bureau of Indian Affairs schools, which NAEP reports separately) and Catholic and non-Catholic private schools is obvious.

However, selection effects--for example, the higher innate student ability or more productive learning environment provided by parents of those who attend private schools--may operate. The statistical attempts to correct for this effect often become so complex, fragile, and contradictory that one can find evidence for almost any specific educational proposition one wishes to make. Neal (2002) summarizes a substantial literature indicating that private schooling raises achievement after correcting for selection effects. The work summarized in Jimenez, Lockheed, and Paqueo (1991), which studies the less politically charged issue of education in developing countries, also overwhelmingly supports the superior performance of private education there after correcting for selection effects. But McEwan (2001) uses Chilean data--Chile being a country with an extensive voucher system--to suggest that selection effects may explain most of the achievement gap. However, Chile (like other countries) substantially limits the ability of private schools to deviate in spending, curriculum (i.e., output) and factor ratios from what prevails in public schools (Merrifield 2005). The manner in which public and private schools compete in Chile is similar to what we would expect if a state steel monopoly were opened to competition, but with competitors required to adhere to the same labor contracts and charge the same prices for the same products. The benefits of competition in such an environment are presumably substantially eroded.

With that caveat in mind, a relatively straightforward technique exists for at least crudely correcting for some of these selection effects. Table 3 shows the performance of the NAEP measures of reading, mathematics, science, and total scores broken down by several characteristics that might result in selection advantages for private schools while not affecting their intrinsic performance--parents' education, race, region, and eligibility for the federal school-lunch program (a proxy for poverty). One consistent pattern emerges--public schools perform significantly worse than private ones across all categories. This result is robust across all different methods of slicing up the data. There is, in contrast, no discernible difference between Catholic and non-Catholic private schools. These data do not conclusively correct for the possibility that other undetectable selection effects operate within the homes of parents likely to...