The most opaque elements of the USNWR law school rankings relate to expenditures. They are 10% of the rankings. USNWR does not disclose them. ABA does not disclose comparable figures. You can access them if you purchase access to the “Academic Insights” platform, but they are not otherwise publicly available (or available for any public distribution).

Nevertheless, they are probably the most significant factor as they relate to rankings. The gap between the top end and bottom end is significant. That gives schools the most opportunity to outperform others and climb in the weighted, scaled scores. They are also the least sticky. While peer reputation is notoriously sticky, and a whopping 25% of the ranking, it contributes very little volatility.

(This is not a new concern. Professor Brian Leiter and Professor Corey Rayburn Yung are just two recent examples of law faculty pointing to these problems.)

USNWR collects information from law schools on “expenditures.” (More on this undefined word in a moment.) It asks for “Instructional salaries,” “Administrative and student services salaries,” “Library salaries,” “Other salaries not included elsewhere,” “Fringe benefits,” “Law school expenses (exclude library),” “Financial aid,” and “Library operations.” There is a separate line for “indirect expenditures and overhead.” Law schools are instructed to “exclude expenditures for the LLM program.” That’s about all the instruction law schools are given.

The methodology reports that they calculate per-student metrics: “The average spending on instruction, library and supporting services (0.09) and the average spending on all other items, including financial aid (0.01): The faculty resources calculation for instruction, library and supporting services is adjusted for cost of living variations in law school salaries between school geographic locations by using publicly available Bureau of Economic Analysis Regional Price Parities index data.”

Some categories seems fairly straightforward. “Instructional salaries,” for instance.

But “Law school expenses,” an open-ended catch-all, remains a decidedly underdefined category, and a crucial (often large) component of the weight of the rankings worth 9%. Other “overhead” and adjacent expenses could fall into the “indirect” expenditures, which are weighted at just 1% of the rankings. The scandal exposed in 2005 revealed everything from water bills to the “market value” of LexisNexis accounts have been stuffed into these metrics. (And there’s a whole separate and complicated issue about how law schools “count” their physical building in these “expenditure” metrics.)

One might think that the word “expenditures” means dollars out the door, but it’s been obvious for years that accounting methods (including depreciation or other cost-basis allocation) have been a part of how schools calculate “expenditures.”

But many law schools report extraordinary expenditures. How extraordinary?

Imagine we say that a law school could calculate X dollars in revenue from “gross tuition.” That is, assuming no scholarships and no “discount rate” (which is a totally implausible assumption, I note), let’s figure out how much money a law school might collect from total tuition from its students.

Many schools report “expenditures” of 2X that number. And some even higher.

How is that even possible? Let’s try another small experiment. Suppose we calculate this in a real dollar total. If a school has, say, 600 students and $50,000/year tuition, we can say that “gross tuition” revenue is $30 million. (Again, an implausible assumption, but work with me.) A school reporting expenditures of $60 million, or $75 million, would be burning an extraordinary amount of additional, outside capital each year if it was truly “expenses.”

Law schools are usually cagey about revealing their endowments. We have some figures at the high end: Harvard Law in 2008 reported an endowment of $1.7 billion, which at 5% would spin off $85 million a year. (Of course, Harvard has around 1700 JD students, not 600.) Yale Law in 2009 reported an endowment of $1.2 billion, which would spin off around $60 million.

But most endowments are much more modest. Emory Law in 2016 reported an endowment of $43 million that would spin off $1.7 million a year—and most of that went toward scholarships (i.e., the 1% category, not the 9% category). The University of Texas Law Foundation presently reports an endowment of more than $150 million, which would spin off around $7.5 million. (At any of these institutions, parts of the endowment may be held by the parent university for use at the law school, so the figures could be much larger.)

It’s possible, of course, that many law schools earn a tremendous amount in one-time or annual gifts, among other outlets. And it’s also possible that parent universities are much more heavily subsidizing their law schools than they’ve otherwise been letting on. Finally, it’s also possible that law schools with rising non-JD programs are funneling the profits from those programs into the JD program.

These possibilities seem limited. I don’t know many schools with small endowments are consistently pulling in extremely large gifts each year. Parent universities can help subsidize in some cases and in limited circumstances, but even these rarely go beyond a few million dollars on an annual basis in the most extreme cases. And it’s not clear how expansive these non-JD programs are—certainly not so expansive at most universities to subsidize a JD program so heavily.

And, as I mentioned, this is an implausible baseline. Most schools have a fairly significant “discount rate” and offer a significant number of scholarships to students, offsetting tuition and “indirect” expenditures. So schools not only need revenue for these “indirect” expenditures, but then they’re also spending far more than whatever the X, 2X, etc. “gross tuition” baseline is.

In short, then, almost however you look at it, it’s impossible to account for law school expenditures under the USNWR rankings, except that there are accounting assumptions being made that bear no relation to the reality of the quality fo the education students receive.

This grows more obvious and pronounced each year. But, like many critiques of the rankings, it seems unlikely to go anywhere anytime soon.