Education reporters such as Chalkbeat’s Matt Barnum continue to cling to the idea that pouring exorbitant resources into an inefficient school system can make a sustainable difference in the lives of America’s children. To support the claim, Barnum points to a couple of recent studies examining the association between court-ordered education spending increases and student outcomes.
Jackson, Johnson, and Perisco (2016) conclude that an annual 10% increase in per pupil spending for all 12 years of schooling leads to an increase of about a third of a year of completed education. Similarly, Lafortune, Rothstein, and Schanzenbach (2016) find that court-ordered spending increases improve test scores for the least-advantaged students by a little under a hundredth of a standard deviation per year.
However, both of these studies suffer from important methodological issues that limit their ability to identify a strong causal relationship between education dollars and student outcomes.
Obviously, court-ordered spending reforms are not random events, so using them to predict educational expenditures still results in biased estimates. The public’s desire to improve education in a given location likely leads simultaneously to court-ordered spending increases and political pressures to improve school quality.
Perhaps more concerning is that event studies like the ones Barnum cites capture an entire package of educational reforms during a particular period. For some reason, authors of these types of studies have chosen to point to spending as the cause of the altered outcomes. However, other reforms such as testing accountability, pay-for-performance, and educational choice happened during the study timeframes.
Costs vs. Benefits
For the sake of argument, let’s assume that the detected effects could actually be attributed to educational dollars.
The study by Lafortune, Rothstein, and Schanzenbach finds that a large court-ordered spending increase only raises test scores by seven-thousandths of a standard deviation per year. For the most-advantaged students, the effects are zero. The fact that the detected impacts are trivial suggests that authors are simply picking up the bias generated by their empirical techniques.
However, even if court-ordered spending decisions were indeed random, another important question arises: do the benefits justify the additional costs? In theory, pouring billions of educational dollars into systems designed to improve test scores should do just that, even if schools are not particularly efficient. Certainly, if researchers found that a million dollars per student per year increased their test scores by a couple of points, we would be neither surprised nor enthusiastic.
Further, even if traditional public schools were efficient in allocating resources towards student achievement, it still is not clear whether test scores actually matter. Allocating more resources towards increasing test scores may actually harm students if educators consequently focus less on molding nearly unmeasurable character skills such as morality and determination.
If we really want educational dollars to matter, we ought to allow children to take their public funds to the schools that are best for them. Only in that scenario will educational institutions have a strong incentive to shape the skills that individual children actually need for lifelong success.