“The question we ask today is not whether our government is too big or too small, but whether it works... Where the answer is yes, we intend to move forward. Where the answer is no, programs will end. And those of us who manage the public’s dollars will be held to account — to spend wisely, reform bad habits, and do our business in the light of day — because only then can we restore the vital trust between a people and their government.”
In his inaugural address, back when there was some hope (however naïve) in the idea of a “post-partisan” presidency, Barack Obama outlined a straightforward doctrine for better government. What works will be expanded, what doesn’t will be ended — no ideology, no unseemly horse-trading, just clean, technocratic utilitarianism, doing the work of the people humbly and efficiently.
For those familiar with the D.C. OSP, this credo looks like just one more item to be added to the administration’s growing heap of broken promises, failed aspirations, and betrayed principles.
What is the D.C. OSP? In the words of Dr. Patrick J. Wolf, the principal investigator of the U.S. Department of Education’s D.C. OSP evaluation team, D.C. OSP is, “...the most effective education policy evaluated by the federal government’s official education research arm.” It is a quintessential example of one of the well-functioning programs that Obama’s address spoke of.
More specifically, the D.C. OSP is the District of Columbia Opportunity Scholarship Program, a federally funded school voucher program targeted at the poorest and most at-risk children in the District of Columbia’s school system. And as the federal budget appropriations process for fiscal year 2011 grinds its way through Congress, it appears more and more likely that D.C. OSP is not merely going to be denied new enrollees and wound down, as the president’s February budget proposal outlined, but instead will be completely axed, with its existing students cast back into the broken system from which they had temporarily escaped.
If this happens, it will be one of the most shameful victories by a special interest group over the general good that has occurred under this administration.
D.C. OSP costs a paltry $14 million per year to fund. Families with an income less than 185 percent of the poverty line are eligible to apply. Those that meet this requirement (as well as a few sensible others, such as residing in the District of Columbia), are randomly assigned — two thirds to a treatment group, and one third to a control group. The treatment group is offered a school voucher redeemable for up to $7,500 at any OSP eligible private school. Each year, approximately 1800 students use a voucher to attend a private school.
In evaluating the performance of this program, there are two important questions: relative to alternatives, what is the cost of program and what is the difference in educational outcomes?
In answering the first question, it is important to make a distinction between marginal and infra-marginal effects. The marginal effect of a school voucher program would be the take-up of the voucher by students who would not otherwise attend private school — the infra-marginal effect is the use of the voucher by students who were going to attend regardless of the incentive. School vouchers are almost always designed to provide a subsidy that is less than or equal to the per-pupil cost of educating a student in the public system. If there were no infra-marginal effect, it would be impossible for the take-up of a voucher to incur a larger government expense than the cost that would have been incurred absent a voucher. However, if most of the voucher users were going to attend private school regardless of the voucher, then it might be possible, as many opponents claim, that the voucher would take funding away from public schools.
What were the results in D.C?
Of the pool of eligible OSP applicants, roughly 75 percent come from the public school system, and 25 percent from charter schools. For students in the treatment group (who are offered the voucher), 19.1 percent choose to remain in public schools, 9.3 percent remain in charter schools, and 71.6 percent enroll in private school. For students in the control group, 53.9 percent remain in public school, 33.8 percent remain in or switch to a charter school, and 12.3 percent enroll in private school.
In rough numbers, D.C. has 50,000 students in public schools and another 25,000 in charter schools. Publicly, the D.C. school system reports that it spends $15,000 per student — however, when one adds in supplemental educational funding, facilities costs, federal funding, teacher retirement benefits, and separates charter-school-related costs from public-school-related costs, the tale is much different. Charter schools cost the system $12,000 per pupil, while public schools cost $26,500 per pupil. If special needs or “private placement” students are removed from the tally (D.C. has 2,400 such students at a cost of $89,500 per pupil), the per pupil cost for the general public school population goes down to $23,250.
In contrast, the average voucher amount was a paltry $6,620.
Assuming that the students included in the D.C. OSP program are statistically average in terms of their cost to the educational system (and adjusting for some minor statistical differences between the treatment and control groups) the treatment group, on a per-pupil basis, costs $6,500 per year less to educate than their control group peers.
With savings like these, one might expect massive decreases in performance. After all, consider the prognostications of teachers unions in New Jersey, where per-pupil spending is roughly on par with that of D.C. Faced with an unprecedented budget deficit, the recently elected Republican governor, Chris Christie, cut $880 million from the state’s K-12 system. Relatively speaking, the educational system got off very light-- $10.7 billion was cut in total, and the state educational system — which at $11.1b represented more than a quarter of all state spending before the cut — was barely scraped. Nevertheless, the teachers unions are in an uproar. Gone is $880 million, and New Jersey, we are told, will sink into the Atlantic.
Imagine for a moment — what if the entire K-12 system in New Jersey had been replaced by a voucher system akin to the D.C. OSP, and had achieved a similar result, namely a decrease of $6,500 in per-pupil spending? At nearly 1.4 million New Jersey K-12 students, Christie wouldn’t have just cut $880 million — he would have cut the entire state funding of K-12! That’s right: Federal and local funds would be sufficient to run the entire system, vouchers and all!
But if the loss of $880 million was a crushing blow, surely a cut that is 10 times as large would devastate the children of New Jersey, right?
Well, what was the experience in D.C?
As it turns out, students who are offered the voucher perform much better than their control group peers. Graduation rates are 82 percent compared to the control group’s 70 percent (and better than the overall system, which manages less than 50 percent). Parent satisfaction is up, as are math and reading test scores (although only the reading test score improvement is statistically significant). There are virtually no negative effects to report — although just applying for the voucher indicates a level of parental and student effort that promises much higher educational outcomes than the city average, it is clear that actually receiving the voucher matters quite a bit as well.
These results are even more impressive when one considers that 25 percent of the treatment group has never used the voucher, and only 41 percent have used it every year it was available to them. Of the group that are offered the voucher and use it, the educational outcomes (as well as, obviously, the cost outcomes) are even better — for example, the graduation rate among students who receive AND use the voucher is 91 percent — a 21-point improvement over the control group and a nine-point improvement over the treatment group average.
Of course, it is statistically invalid to exclude from the analysis those families who are offered but do not take up the voucher. To do so could introduce selection effects into the data — in other words we would not be sure that difference in outcomes between the treatment and control group are due to the efficacy of the treatment or merely attrition of the lowest performers from the treatment group.
But there is ample reason to believe that there is no selection effect at work here, and instead that the true effect is as large as the audited statistics suggest. Firstly, there is great statistical similarity between those in the treatment group that did not use the voucher and the members of the control group in terms of background and their ratio of public to charter school attendance. Secondly, the most common reported reason why the voucher is not used is that there are not enough seats open at the private school of choice. As private enrollment surged, demand has simply outstripped supply, and a lag in the market has denied some treatment subjects the opportunity to exercise the voucher and experience its benefits. If D.C. OSP were continued, presumably the market would adjust to this influx of customers, and the observed benefits would be even larger.
There is plenty of room to debate shades of success and whether or not D.C. OSP would show equally strong results if implemented across a broader section of the population. But nowhere in the data is there a legitimate case for ending the program altogether.
In short, it is special interests, namely teacher unions, behind the premature end of the D.C. OSP. This is understandable — they have their sinecures, and they wish to protect them from market competition. But even so, consider the context. D.C. OSP is free money for the educational system. Not only does it improve public school resources indirectly, by reducing the number of students that need to be served, but it also does so directly, by throwing tens of millions at the D.C. system as a pork-barrel sweetener. The unions have deemed this program such a threat that they are willing to cut their own public school funding if thats what it takes.
This is more than just bad government. Ninety percent of the beneficiaries of the D.C. OSP are black and nine percent are Hispanic. These are not the well-heeled elite of nearby Tyson’s Corner. These are the downtrodden underclass, our inner-city equivalent to the waves of boat people that found our shores in the wake of the Cuban revolution or the fall of Saigon — desperate families seeking asylum from a collapsing system. If the rhetoric of the Democrats were matched by action, and if the stereotypes of the Republicans are believed, these people would be the natural constituency of Obama’s party — indeed, he would likely point to them and their plight as the very problem that inspired him to enter politics. And yet, there is no Operation Pedro Pan or Babylift on the horizon for D.C.’s poor. Having fought loyally by Obama’s side during the campaign, those D.C. residents struggling at the margins of society are to be rewarded by being abandoned to the fates. In a foreign policy context, this sort of betrayal of former allies would be considered rank cowardice, it would be an indictment of our moral leadership, it would burn on our conscience as a nation — it would be as black a mark on our history as when we turned away Jewish refugees from our shores.
Unfortunately, in the sausage-making of this administration, it rates just as business as usual.