Education Next has a new story by David Bass about possible fraud in the National School Lunch Program (NSLP). As EdNext is wont to do, the article leads with a provocative graph, this one purporting to show the rise in free and reduced-price lunches as the percentage of children in poverty has remained more or less constant. By using the percentage of school lunches served free or at reduced price, their graph is incredibly misleading, and it buries Bass’ main point–that the income verification measures in the NSLP are weak or nonexistent.
A July 2008 review of the NSLP by the Department of Agriculture found what the EdNext graph shows–that, of lunches served, a higher percentage are indeed free or reduced-price.
But that’s not the full story. What EdNext omitted is that the Department of Agriculture report also looked at student participation rates, not just as a percentage of lunches but also as a percentage of kids, and they saw relative stability:
The rate of program participation has stayed fairly stable since 1989, the earliest year for which data are available. In Federal fiscal year 1989, NSLP participating students (who received a full price, reduced-price, or free lunch) accounted for 60 percent of all students in NSLP-participating schools. The rate declined slightly through the 1990s to nearly 58 percent in fiscal year 2000 and then increased steadily from 2003 to about 62 percent in fiscal year 2008, the last year for which data are available.
In other words, participation in the National School Lunch Program has fluctuated between 58 and 62 percent over the last 20 years.
The text of Bass’ piece explains how the NSLP has poor mechanisms for verifying family income, and that this loophole leaves the program open to fraud. This has all sorts of consequences ranging from costing federal taxpayers money to negatively impacting research on schools (the free or reduced-price lunch figure is often used as a proxy for low-income students) to harming school choice assignment plans that use income as a factor. Bass may or may not be right on the merits (his claim that large percentages of unreturned verification forms indicates widespread fraud seems like a stretch), but EdNext did Bass no favors by pairing his article with such a misleading graph.