Vouchers? A bad idea and here are several reasons why.

Posted on May 18, 2011


This post reviews a short paper by Dr. John Merrifield and Joseph Blast as well as research on vouchers in peer-reviewed journal articles and books. The paper by Merrifield and Blast has not been peer-reviewed or published in any manner other than a policy brief. I’ve only had 5 hours to work on this brief (thanks to Rep. Miller who did not even have the guts to bring a bill to a committee, but is taking the easy way out and amending to another bill) and will be adding citations and other information throughout the day.

For those of you who don’t have the time to read the entire post, here is a summary of the points, most of which are detailed in the post and ALL of which are supported by evidence or research:


The paper by Merrifield and Blast has serious flaws that make their conclusions difficult to accept. These flaws include reliance on ONE STUDY that they fail to adequately describe, faulty data (that easily could have been corrected with a few hours of work), and assumption laid upon assumption that is not supported by any research or by apples-to-oranges comparisons that don’t provide support for a statewide program. Thus, the quality of the paper is low and should be read with extreme caution given the errors and assumptions inherent in the paper.


Private school tuition is far greater than what the authors claim to the point that the voucher would not cover most of the urban area tuition costs for private schools and certainly not all of the associated costs with enrolling a child in a private school (such as uniforms, fees charged by the school that can be up to $1,000 per year, transportation, and time to provide transportation).

3) RESEARCH SHOWS VOUCHERS SEGREGATE STUDENTS. Vouchers increase the segregation of students by race, class, and ability.

4) VOUCHERS DO NOT IMPROVE ACHIEVEMENT. Vouchers do nothing to improve achievement except for African American males in urban settings, but not even in all urban settings.

5) STATE TAX DOLLARS WOULD SUPPORT ANY PRIVATE SCHOOL, INCLUDING ALL RELIGIOUS SCHOOLS. State tax dollars would be supporting all types of religious schools: Catholic, Baptist, Episcopalian, Jewish, Muslim, etc.

6) THERE WILL BE NO ACCOUNTABILITY FOR SUCH SCHOOLS TO USE OUR TAX DOLLARS WELL. Students will not have to take any type of assessments, there will be no state oversight of the schools, and there will be no data collection on the students or school. We will be handing over tax dollars and saying, “Here–take our money, we trust you.”

7) SCHOOLS CAN DISCRIMINATE AGAINST CERTAIN TYPES OF APPLICANTS. Private schools are not required to take all students. Most require an application from the family and require good grades, a history of good behavior, and many require students to pass a standardized test to get into the school. Bilingual and special education students will largely be excluded as well as those who struggle academically.


Costs of Voucher Plans

In their paper, the authors provide the following data. Their source for the Texas Parochial Schools is the Archdiocese of San Antonio. Evidently no attempt was made to verify the numbers provided by the Archdiocese. More importantly, the authors did not obtain tuition for parochial schools in other areas of Texas or for independent private schools anywhere in Texas.

Table 1: Average Private School Tuition (Adapted from Merrifield & Blast, 2011)

In just a few minutes conducting a web search, I was able to find a website that linked to private schools in all areas of Texas that included enrollment figures and links to individual school websites.See http://www.privateschoolreview.com/

Given that I only had a few hours to prepare this post (thanks to Representative Sid Miller (R) of Stephenville who will be proposing a school voucher amendment today without having the strength of character to author a bill and have it debated in committee).

Let’s see what ACTUAL SCHOOL WEBSITES listed for tuition in San Antonio Catholic Schools

Table 2: Tuition and Tuition and Fees for Parishioners and Non-Parishioners in San Antonio Catholic Schools

Clearly, the average tuition is greater than the tuition figures provided by the Archdiocese and the figures used by the authors. If we cannot trust the authors to do some RESEARCH to determine some accurate numbers for the entire foundation of their paper, how can we trust them on the rest of the assumptions and calculations?

Note that the authors do not discuss any of the associated fees built into the direct cost of enrolling a child in a parochial school. Each school had a variety of fees that totaled anywhere from a few hundred dollars to almost $1,000. In addition, the tuition and fees on the websites did NOT include the costs of buying uniforms or providing transportation to and from the school. Given the price of gas and the cost of the time spent driving a child to school, transportation costs could be a major expense. Finally, the cost of living in San Antonio is lower than in the other major metro areas that enroll substantial numbers of students (Houston, Dallas, and Austin). Thus, the parochial fees are greater in those areas than in San Antonio).

The authors also claim that Catholic schools enroll more than all other religious schools combined. This is true, but just barely as shown on page 7 of the document by the National Center for Education Statistics that can be found at: http://nces.ed.gov/pubs2009/2009313.pdf.

The authors also relied on only one study conducted in 1990 in a relatively obscure journal (not available on-line at The University of Texas library which has one of the largest on-line journal collections in the world). The authors do not tell us anything about the sample or the location of the sample, the actual methods or the actual results. Thus, we don’t really know anything about the details of the study. For example, the authors do not report the odds ratio for the logistic regression analysis, so we don’t know whether the finding was strong or weak. If weak, then the authors’’ entire premise rests on a very, very shaky foundation. Given that one of the authors is an economics professor, he should have known to include this information somewhere in the report.

Further, the 1990s study did not include the effects of charter schools which research has found to have a NEGATIVE effect on private school enrollment. Thus, the estimate used from this study is clearly inaccurate.

And finally, every estimate in a regression analysis has error. Thus, the 5% increase in enrollment the authors’ claim from the study could be lower of higher. How much lower or higher we don’t know because the authors did not provide that data.

Faulty Assumption about Tuition Elsewhere in Texas

As shown in Table 3, the voucher would not cover tuition at ANY of the most popular private schools n Houston. Nor would it cover Catholic school tuition in Houston. In fact, the price differential between tuition (excluding fees, uniforms, transportation, and time) is so large that most families could not afford tuition even if $5,000 was subtracted from the current tuition.

Table 3: Tuition in Houston Area Private Schools with the Largest Enrollment

For Dallas area schools, see: http://dallas-area-schools.blogspot.com/2011/04/private-schools-ranked-by-cost-of.html  Note that the voucher would cover ONE school on the list although the list is not exhaustive.

Tuition rarely covers the true costs of operating religious schools or even private, independent schools (Baker, 2009). Indeed, perusing the various school websites of private religious and independent schools in Houston, San Antonio, and Dallas reveal that the schools generate income from donations from private individuals and corporations. So, for example, if the state provides $5,000 for another 50 students to enroll in a particular school that charges $5,000 in tuition but has actual operating expenses of $7,000, then the school would have to increase non-tuition revenues by $100,000 or increase tuition. Thus, as Baker (2009, p. 10) asserts, such an analysis “suffers from the central problem of asserting that private schools can take on additional students at then existing tuition levels and subsidize the difference via philanthropy.” Hence, either tuition is going to have to increase far more than the authors assume or philanthropy to private schools will have to increase far more than the authors assume.

Apples-to-Oranges Estimates

The authors use the aforementioned study and two urban area studies to identify estimates of how many students would enroll in private schools. Yet, isolated programs and statewide programs would have different effects on enrollment in schools of choice. Yet, the authors never mention this.

Faulty Assumptions about Supply

The authors assume that the inputs needed for schooling are not scarce. Yet, almost every school in Texas has a shortage of math and science teachers. Further, given that non-elite private schools have lower teacher pay, the schools participating in this program would likely have extreme difficulty in attracting and retaining high-quality math and science teachers. While they note 200,000 teachers are produced each year, they fail to mention that is a national figure. The total is closer to 20,000 in Texas.

On page 12, the authors assume that any increase in tuition would not affect the choice of parents, but this assumption is based on tuition being less than the voucher. Yet, as I showed above, the vouchers very likely will not cover most tuition costs at present and the voucher certainly would not cover tuition if tuition increased due to greater demand for seats in private schools.

Also on page 12, the authors assume that competition increases the efficiency with which resources are used. However, they provide no evidence that this is the case in education nor do they note that competition does NOT improve student outcomes.

The authors assume that there would be negligible tuition increases and they cite the Milwaukee and Edgewood experiences as evidence of this. But again, there are assuming that the results from single-district experiments would be the same as for a statewide program.

Faulty Analysis of the Impact on State Taxpayers

The authors use data from 2008 in determining the cost of those already in private schools returning to public schools for a year and then returning back to private schools. Thus, their estimate is too low. Further, they assume—without a shred of evidence—that only 6 to 7% of families who currently enroll in private schools will send their children to public schools and then return them to private schools.

What if rather than 6 0r 7% transfer, 20% actually transfer? Or 50%?  Then the cost estimate starts to change significantly.

Further, note that the authors totally ignored the cost of administering such a program. The administration costs would be rather large due to the initial start-up costs and the need to ensure against fraud.

Vouchers subsidize wealthier families’ choice to send children to private schools

The vast majority of those in private schools–especially non-religious/independent schools–are from families that are not poor. For example, see Table 4 below from the National Center for Education Statistics.

Table 4: Characteristics of Students in Different Types of Schools

While the authors argue that this is the very point of providing a voucher, the voucher of just over $5,000 would not cover full tuition at most schools and leave parents with at least an additional $2,000 in extra costs in the cheapest schools as much as $15,000 in the most exclusive private schools.

Restricted Enrollment in Private Schools

Unlike public schools, not every student would be eligible to enroll in private schools. Every independent private school I reviewed had fairly stringent admission requirements that typically included, at a minimum, a review of student grades, student behavior records, and standardized test scores. Further, applicants were required to read the application process that was written only in English and were also required to produce a social security card. This application process would prove to be a significant barrier to families whose parents do not speak or read English. Likewise, most private schools do not enroll profoundly disabled students (who are extremely costly to educate) or even moderately disabled students. Most of the private independent schools cautioned applicants that the school held high standards and no special assistance would be provided students who could not meet the rigorous standards.

Segregation of Students

There is ample research evidence (Ladd, 2002) that finds school choice programs segregate students on the basis of race, ethnicity, socio-economic status, and ability level.

Vouchers do NOT improve student achievement

Reviews of the research on the relationship between school voucher plans and student outcomes by Baker (2009) and Ladd (2002) find the following:

–Increased graduation rates

–No effect on achievement except for African American students in inner-city schools.

–The positive effects for African American students are small and inconsistent across settings.

–Students transferring schools has a negative impact on student achievement, thus vouchers have an initial negative effect on achievement that is overcome only if a student remains in the same school;

–There is a profound negative effect on achievement for students opting out of public schools for private schools, especially for students close to the terminal grade level of their particular level of schooling (Özek, 2009).

Posted in: Uncategorized