How much should law school cost?

Consider the following radical thought experiment: How much would law school cost if that cost were determined by what a law degree is currently worth?  To answer that question, I'm going to make the following assumptions, both of which flow from conventional analyses of what constitutes a reasonable relationship between cost and benefit when investing in educational credentials:

(1)  The  median educational debt acquired by law students should not exceed their median starting salaries.

(2)  The relationship between advertised tuition and loans taken out is highly predictable.

The median starting salary of graduates in the class of 2011 was, optimistically speaking, around $45,000 (NALP lists a median of $60,000, but this is based on salaries for just 42% of the class, and given that there's a tight correlation between higher salaries and the likelihood they're reported, and that nearly 15% of the class had no salary at all, an estimate of $45,000 is if anything generous).

But let's give law schools every benefit of the doubt and assume it was $50,000. Let's go further, and not take any other educational debt into account when estimating law graduate debt.  What would tuition have to be to generate a median law school debt at repayment of $50,000?

Note that every single law school in the country that accurately reported mean amounts borrowed during law school by its class of 2011 reported a higher total than this.  (Drexel, Southern, Texas Southern, and Georgia State initially reported lower numbers to the ABA, but their reports are obviously for one year of debt rather than three.  I don't know whether they've corrected these figures with the ABA, but they remain uncorrected on USNWR's webpage, which is the only public source for this information).

Note too that borrowing $14,600 per year will, because of accruing interest, result in a debt level of $50,000 in law school debt alone at repayment.

For the mean law school debt -- the median would of course be higher -- of a graduating class to be in the $50,000 range, tuition should be no higher than $10,000 per year.  This can be deduced by comparing historical tuition levels to amounts borrowed (This assumption is generous to schools located in expensive cities.  In such cities, a student who debt-finances a $10,000 per year tuition bill would have to pay cash for almost all of his or her living expenses during three years of law school to avoid incurring debt at graduation of more than $50,000. Even in areas with a modest cost of living the large majority of a student's living expenses can't be debt financed if the student paying a $10,000 tuition bill is going to graduate with not much more than $50,000 in law school debt).

Now it's a sign of how warped the economics of law school have become that at almost all law schools the idea of charging $10,000 per year in tuition has come to seem almost inconceivable.  But in fact there's absolutely no reason why law schools couldn't be operated perfectly well while charging that amount.

Here's CU's annual resident tuition IN 2011 DOLLARS:

1981:  $2,700

1991:  $6,000

2001:  $8,300

Naturally these tuition levels were made possible in part by tax subsidies.  But note that even if tuition in 1981 had been 100% tax-subsidized (in fact it was nowhere close to that) the actual cost of educating each graduate was still hardly more than $10,000 per year.  By 2001 the law school was getting less than a third of its operating budget from the state, so the actual per-student cost of operation was still not much more than $10,000 (since all non-resident CU students get resident tuition after their first year, out of state tuition has very little effect on this analysis).   In fact, Harvard Law School was charging only $15,862 per year in 1981 in 2011 dollars, and even as late as 1985, i.e., several years after the beginning of the big tuition run-up, median tuition at private ABA law schools was still only about $15,000 (resident tuition at public law schools was under $4,000).

And none of this takes into account the many ways in which technology should be reducing the cost of education.  Adjusted for inflation, I paid more for a Honda Accord in 1995 than I would pay for the -- much superior -- 2012 version of that car.  The computer I'm typing this on couldn't be sold for $100 (it's four years old), but it still has more computing power than machines that cost several million dollars thirty years ago.  Why should education be different?

One answer people sometimes trot out is the Baumol effect.  But the very example Baumol and Bowen use to illustrate their theory actually illustrates its inadequacies when it's deployed to explain why tuition is so high:

Baumol and Bowen pointed out that the same number of musicians is needed to play a Beethoven string quartet today as was needed in the 19th century; that is, the productivity of classical music performance has not increased.

This line of analysis fails to take into account that the cost to consumers of a performance of a Beethoven string quartet has declined by a factor of close to infinity since Beethoven's time.  The relevance of that insight to the consumption of higher education, in a world where the cost of transmitting knowledge is hurtling toward zero, should be obvious.

The only reason law schools charge $30,000 and $40,000, and $55,000 per year in tuition is because they can.  Those prices bear absolutely no relationship to either what a law degree ought to cost to acquire, or to what it will be worth to those who acquire it.