[This post has been updated as noted in brackets and italics below to account for responses to it and to attempt to help clear up some of the confusion out there about valuation methods.]
In the world of corporate finance there is a very simple decision rule: managers should accept all positive net present value (NPV) projects. This is a very simple and powerful concept. It lies at the heart of how we train managers at every business school in the world.
The steps to apply the rule, however, can be deceptive to the uninitiated.
Thus, it is perhaps understandable that Paul Campos, the notorious “bad cop” of law school reform according to the Cato Institute, who is trained in something called “law and literature,” believes he has debunked the most important finding of a recently posted study of the value of earning a JD called The Economic Value of a JD.
[Update - Professor Campos' confusion about present value continued today in his response to my blog, so it might help him and some readers to take a look at these slides prepared by Campos' Colorado colleague Jaime Zender, a Yale trained economist and endowed chair at CU's Leeds Business School. As Professor Zender states: "The net present value rule states that you should accept projects with a positive NPV and reject those with a negative NPV. All agents agree on the desirability of positive NPV projects."]
The study in question was authored by Frank McIntyre, a Stanford-trained labor economist and finance professor at Rutgers, and Michael Simkovic, a Harvard-trained empirical law and economics scholar who used to work at McKinsey and Davis Polk and is now on the faculty at Seton Hall. It is available here.
The steps in applying the rule require a projection of cash flows over a certain period of time, an estimate of the startup costs for the project and the calculation of an appropriate discount rate. All three of those components are then combined by dividing the future cash flows by the discount rate and then subtracting the startup costs.
(Formally, PV of a lump sum = C/(1 + r)[t] where “t” is an exponent for the time periods. If the startup costs are spread out over time they, too, would need to be discounted back to present value. As I note below, tuition is such a startup cost.)
If the resulting figure is positive, the rule dictates that the project should be undertaken. This is true even if that positive number is very small. It is this last dimension of the rule that can be the most mystifying and it is, indeed, the one that has thrown off Professor Campos.
He argues in a recent post at the equally notorious website “Lawyers, Guns and Money” which Professor Campos uses as his home base (Guardian journo and Edward Snowden interviewer Glenn Greenwald refers to LGM as a “filthy cesspool” and a “cesspool of unprincipled partisan hackdom”) that the discounted value of a JD at the median is something north of $100,000. He concludes that he has therefore succeeded in “critiquing” the paper by Simkovic and McIntyre.
He has, however, done nothing of the sort.
He has simply done exactly what the paper implies all analysts of the value of a JD should do – apply relevant costs such as potential debt and taxes and subtract those from the expected and discounted future cash flows. Sure enough, even with his inputs (like the implication that students must borrow $200,000 to go to law school) the result Professor Campos comes up with is a NPV of $109,000, i.e., positive.
[UPDATED: In fact, reaching that 109K result assumes that he uses the correct inputs, but Campos also includes cost of living in the $200K figure he uses as a borrowing cost when those are (roughly) the same for both JD's and BA holders who do not go to law school. He also includes undergraduate debt which of course a BA holder also has. And at one point he even suggests one should subtract the future value of the 200K (440K) from the earnings premium. That makes no sense - the future value of 440K and the 200K are the same values expressed at different times - that is the entire point of the exercise. Fortunately for all concerned he backs away from that strange exercise. Instead, in what might be called the "alchemy of Paul Campos" he turns 200K of future debt into 311K of present value and deducts it from the accepted 420K PV of future earnings to reach his final figure of 109K. Of course, even if indeed a law student needed 200K on the first day of law school to attend there would be no reason to transform that into 311K.
[Giving Professor Campos the benefit of the doubt I think what he actually did was come up with future value of law school debt and compare it with present value of earnings. That's apples and oranges. You can use future value of debt but then you also have to use the undiscounted value of the lifetime earnings premium which, of course, is far higher than $420,000. That would mean approximately $910,000 post-tax at the median and $1.33 million at the mean post-tax.
[In any case, instead, Campos should have only subtracted the three years of tuition discounted back to present value, as explained by Simkovic and McIntyre, an amount closer to 90K when one includes grants, etc., based on ABA data. The opportunity cost a law student pays because she is not working while a BA is is reflected already in the $420,000 earnings premium that Campos concedes. Tuition is the only material cost a law student pays that a BA who does not go to law school does not. Even Derek Tokarz, a leading figure in the critics' camp, agrees this is appropriate, commenting on Concurring Opinions (#36): "Your study puts the average net cost (sticker minus scholarship) at $90,000, and that looks reasonable even against the 25th percentile's $350,000 premium." Unfortunately, Tokarz then makes the same mistake that Campos made and uses a future value calculation of debt, again comparing his oranges to the authors' apples.
[When you use the correct inputs you end up with $330,000 not $109,000 in PV. How? The median earnings premium for a JD holder in PV terms is $420,000 after tax and if you subtract the PV cost of tuition, $90,000, you end up with $330,000 not $109,000. This is an apples to apples comparison. The mean figure is substantially higher, of course.]
In any case, the central point is that Campos concedes the number is positive.
Now return to what I said at the outset: managers – in this case the managers are prospective law students “managing” their own human capital – should accept a project with positive NPV. In other words, even under Campos’ analysis the conclusion that a student should reach is that law school makes sense. The actual net dollar amount above present value is irrelevant as long as it is positive.
The discount rate is a challenging figure to determine properly (at one point, for example, Professor Campos asks his blog readers for their view on the correct rate, even though he had already posted his conclusion using a supposed rate) but it should take into account all of the quantifiable risks associated with the project. Of course, if a student can find another project that has a larger NPV it would make sense to undertake that project. But Simkovic and McIntyre have already considered that because the cash flows they examine are based on the earnings premium that comes with a JD above and beyond what a college graduate would earn without going to law school.
They also suggest, reasonably enough, that the range of graduate study opportunities for most JD applicants are limited in value. Nearly 70% of law school applicants come out of social sciences and the humanities not STEM or business school (much like Professor Campos and therefore not likely exposed to the subtleties of the valuation process).
Of course, a handful of such students might get lucky and wander into the marketing department writing ad copy for pre-IPO Google….well, one can see the problem.
So, far from undermining the Million Dollar JD Value paper, Professor Campos simply confirms its fundamental insight: law school is a positive net present value project for the vast majority of law students even when tested by the institution’s leading opponent.