Monthly Archives: July 2013

Thoughts on ‘Part 1 first.

C. How Much Should a Prudent Student Be Willing to Pay for a JD?

Brace yourself. The exciting bottom line at the end of all that follows is that it depends on several factors.

Simkovic & McIntyre argue that a JD is underpriced, because the substantial majority of students armed with a JD can expect increased earnings over a BA with a net present value well in excess of the cost of the degree — even at full freight.

So far, I haven’t seen a convincing critique of the core of the study, although I do think there are a number of things at the margin where the authors have been unduly optimistic. For example, they assume law grads get a JD at 25, maximizing the lifetime earning stream; they assume law students earn an average of $24,000 in summer and term-time work during the 3 years of law school, which I think is high. The first is more significant than the second, although neither suffices to do more than put a small dent in the overall findings for most JD applicants. Another issue of note is that the study compares a JD to a no-JD, ie BA-only, world. It doesn’t compare a JD to an MBA or other advanced degree. Even taking all the study’s assumptions on board, the premium for alternate degrees might be higher than a JD.

But it’s one thing to set out some numbers about the distribution of national wages in the past, another to suggest they constitute a trend that will continue in the future, and still quite a different third thing to suggest that Alice and Bob can use these data to make a rational economic choice about how much they should be willing to pay for a JD. The headline numbers in the study are probably these:

Pre-Tax Present Value of JD Premium over BA (prior to calculating cost of tuition) 3% discount rate

source: Table 8

On the one hand, all these numbers are positive, so the prediction is that — even after one further subtracts the cost of tuition — a JD pays off for at least 75% of graduates and probably a lot more. On the other hand, some of these numbers are not that big, notably the 25th percentile number for men and women is pre-tax, so knock off another 30% or so, and more in some states. This starts to look less and less a winning proposition for anyone paying full private-school tuition and falling noticeably below that 25% mark. [Remember, that 25% is not Alice or Bob’s class rank or their relative outcome in their law school class — it’s their relative salary outcome in the national JD class.]

Even if we assume a (perhaps slightly discounted?) version of the authors’ numbers accurately represents the historic earnings premium for a law degree for men and women at the 25th, 50th, and 75th percentile of the earnings spectrum for JD-holders, how can a prospective student today use this information to decide if a law degree is a good personal investment? (I should note here that the calculation is strictly economic; to the extent that the experience of law school and/or being a lawyer is perceived as fun or excruciating, as the case may be, these are external factors to the economic analysis — although they should in no way be external to the decision as to whether to pursue a JD. It is, however, fair to ask whether a law degree will be affordable even if law school or law practice is fun; or whether a law career will pay enough to make up for law school’s or law practice’s misery.)

Here is where the complexities begin. The authors give us a matrix of outcomes for representative men and women at three points on the earnings spectrum. We also know, from well-publicized anecdotes and blogs,¹ if nothing else, that for some number of recent law graduates — the ones with large debt and lousy prospects — that the situation is currently dire. Indeed, it is the plight of these people, who likely number in the tens of thousands, that I think so inspires the anger of people who have been beating on Michael Simkovic and Frank McIntyre’s somewhat smug attempt at a refutation of the claim that the sky is falling.

What are the take-home lessons in ‘The Economic Value of a Law Degree’ for Alice and Bob? In other words how should Alice and Bob think about their potential earnings, and what should that mean for their willingness to shoulder law school expenses?

The paper uses aggregate figures. It doesn’t differentiate by law school. It ranks people by economic outcomes, which are only partly correlated with inputs — how well people did before law school, how prestigious a law school they go to, how well they do in whatever law school they attend. Since no potential 1L can know with great certainty how they will do in law school, nor what sort of legal (or non-legal) career they will find they desire, or be forced to settle for, much less the extent to which that success or failure will translate into income, Alice and Bob like all potential 1Ls must in effect run some simulations based on assumptions about their economic potential and their economic prospects.

The first thing Alice and Bob need to do is to think about what it is they might want to do. Yes, poeple’s ideas about careers often change radically in law school. But then oftentimes they don’t. Is their goal is primarily financial, or is it to be a crusading prosecutor or a defender of DREAMers and other immigrants? Some jobs pay a lot more than others, and if they aspire to one that isn’t highly paid, that may put limits on what they can afford to pay unless they have rich and generous relatives.

The next thing is to make a sober calculation about where one will fall in the national income distribution. This isn’t easy. It’s not a simple matter of plugging in LSATs into some formula based on a law school’s US News rank, and not only because those rankings have only limited meaning. Remember that about 1 in 5 law graduates doing legal work enjoyed even worse outcomes than the 25th percentile reported by Simkovic and McIntyre.² At some point not too far from there — the bottom 10%?15%? — law school becomes an economic loser if you have to pay full freight, and at some point lower down (we have no real idea where but it is certainly there) it becomes economically dubious even if you get a full ride due to the opportunity costs.

If I were advising a student, I’d start with the extremes: if you are contemplating going to a top-three law school, it’s a no-brainer: you’ll have a great intellectual experience, and the odds of paying off whatever you have to pay are pretty good. Get a bit of a discount from a “top 14” school and I think the odds are also excellent that (financially) you are getting a good deal; it’s quite likely this is also true at full freight unless you do (or interview) very badly.

Conversely, if you are going to a very-low-ranked law school — which I might define expansively to include a decent fraction of the law schools outside the USN top 144 — I’d encourage Alice and Bob to assume, unless they have a promise of a job in the family firm or they just need the degree to get a promotion in the organization where they currently work, that you will be in the bottom 25% percent of the distribution. Unless you have some reason to believe that this school has a lock on a strong regional market, beware.

That leaves many people in the middle. In order to make a rational economic calculation Alice and Bob need to make educated guesses about three things, two of which are hard to guess: 1) to what extent the Simkovic & McIntyre numbers should be discounted because of things knowable before starting school (such as starting law school later in life or carrying a lot of high-interest college debt); 2) Where they believe they are likely to fall in the post-JD income distribution, taking account of their taste for risk; 3) Whether the real rate of interest applicable to their circumstances will be more or less than 3%.

That last point is surprisingly important. Consider this back-of-the-envelope recalculation of the chart above. This one uses rounded post-tax numbers, and a 4% real (ie inflation-adjusted) interest rate instead of a 3% rate. This 4% rate more closely reflects what students who graduate with $150,000 of debt and up normally have to pay on the weighted average of their debt.

Present Value of JD Premium over BA (prior to calculating cost of tuition)
Post-Tax at 4% discount [based on combining tables 7, 8 & 9]
(Table 9/Table 7)(Table 8)(1-tax rate, see p. 43)

Now things look considerably more bleak for the students in the bottom quarter of the income distribution. Whether their JD is a good investment is acutely sensitive to how much it costs. A full-freight degree at a private law school will cost upwards of $130,000 (recall that the study assumes living expenses are the same as an alternate choice; it ignores moving costs too). All of a sudden that present value $200,000 (minus summer & term-time earnings, minus other adjustments) doesn’t look so great — and the lower the graduate falls below the 25th percentile on the income distribution, the worse it will get.

Even before that point, life can get very uncomfortable if you have a lot of undergraduate debt, especially in the early years when salaries are low but repayment is hitting as hard as ever. Economists figuring out the present discounted value of a degree ignore cash flow issues because they use a present discounted value that sums the values over a time series. They don’t ask if every year in the series is positive, but only whether the total is positive. Simkovic & McIntyre suggest that cash flow is an issue they can reasonably ignore because so few JD holders default on their student loans. They also say that,

Our estimates of interest rates faced by borrowers are probably too high —that is, the present value of the degree we calculate based on the figures in Table A1 will be too low—because we have not included tax incentives or generous loan forgiveness programs for low income borrowers. (p. 55, footnotes omitted)

Alice and Bob should, I think, be wary of these arguments. Here is where using aggregates hides a great deal of the important variation in the particular, and where the question for each Alice and Bob is the cost and the the value of their degree.³ Even if average debt is manageable for JDs as a class, some people, the ones with the most debt, will experience a serious cash flow issue in the early years after law school.

So the bottom line here is in fact quite predictable: Whether a JD is a good financial bet is a function of institutional prestige, the price of the degree net of grants and scholarships, regional variations in expected outcomes, special factors that alter value or opportunity cost (e.g. age or giving up a highly paid existing job), and law school outcome (career choice, actual honors, and class rank). All but one of these variables are, in theory, knowable in advance, and that one — law school outcome– should be estimated conservatively (see part I on the Dunning-Kruger effect).

In short, whether a JD is a good financial bet depends on all the things any sensible person would already have thought it depends on. The more debt you are taking on, the more cautiously you should weigh your options. The better the school you are going to, the lower the fees you are being charged, the more law school looks like a good bet. And you need to take a very realistic, even pessimistic, look at where you likely will fall in the class at the school of your choice. On the other hand, the suggestion that law school is a bad economic bet is — if Simkovic & McIntyre’s numbers hold up — likely not true for a substantial majority of would-be JD’s…but by no means all.

D. Policy Implications

Three policy implications flow from the above.

First, we don’t need any more bad law schools; indeed unless something changes to increase the market for legal services, from a strictly economic perspective (ignoring the possible social value of having more lawyers around to serve under-served populations’ needs) we might be better off with somewhat fewer.

Second, law school transparency is really important. Alice and Bob can only make good decisions about whether to sink a massive investment of time and money into a JD if the schools seeking their matriculation provide honest and easily comparable data about student outcomes.

Third, doing something about the undergraduate student debt crisis — such as returning to the norm of a free, or nearly free, public undergraduate education, would not only serve important goals of producing an educated and aware citizenry, but would have a knock-on effect on the JD valuation issue. The combination of high undergraduate debt with high law school debt may be particularly difficult to justify unless the student is acquiring a relatively valuable JD and/or is paying relatively little in tuition.

1. I once wrote an article entitled, The Plural of Anecdote is Blog [↩]
2. You may be wondering why I said “1 in 5” rather than “1 in 4” which would be what you would expect for a number that is labeled the “25th percentile”. This is why: as Simkovic explained,

   for technical reasons related to regression of earnings to the median, our 75th and 25th percentile values are probably too extreme. The “75th percentile” value is likely closer to the 80th or 85th percentile for lifetime earnings, and the “25th percentile” is likely closer to the 20th or 15th percentile.

   [↩]

3. I am also wary of the argument that the low default rate by lawyers is as significant a finding as it sounds. Lawyers are more likely than others to understand how difficult it is to discharge student loan debt in bankruptcy. They might be concerned that defaulting on a loan might reflect on their bar membership or that the publicity might deter future clients or employers. [↩]

Some Thoughts on ‘The Economic Value of a Law Degree’. Having read a great deal of the commentary on Michael Simkovic and Frank McIntyre’s paper The Economic Value of a Law Degree, as well as the paper itself, I can’t resist making some more comments of my own.

It seems to me that a fair amount of the controversy about this paper is that it does one thing (report on the recent history and distribution of the present discounted value of the earning premium from a law degree as opposed to no law degree), then suggests how its results can or should be used to do a different thing (decide if law school is a good financial bet today). It’s easy to misunderstand the relation between the two things. Some critics misunderstood the relationship between the two things; some others didn’t, but think that the authors of the original paper are (negligently or strategically) inviting prospective law students to misunderstand that relationship.¹

In part 2 of this analysis I will suggest that a significant part of the dispute here also can be traced to a different understanding of what real rate of interest ought to apply to a prospective JD’s calculations. Simkovic and McIntyre use a 3% rate, which may in fact be appropriate for the average student. But you can drown in a river that is an average of six inches deep. The more debt students carry, the higher the interest rate they tend to have to pay. Thus for students with a lot of college debt who also borrow heavily for law school, the real rate is 4% or even more. (See the very last table in the paper, Table, Appendix Table A-2.) For a significant minority of these high-debt students — perhaps almost a quarter of the students carrying $200,000 or more debt upon law school graduation — a JD may not in fact be a good investment.

The other big source of controversy has to do with radically different visions of the likely future. Those who believe that the law is undergoing a structural transformation (e.g. outsourcing, mechanization, permanent change in customer demand, or inelasticity of customer demand, especially at the high end) think that any backward-looking study is a poor guide to the future. The authors’ response is that any structural transformation in the provision of legal services is likely to be mirrored by similar changes in many other careers open to BA-holders contemplating a JD. They argue that their study, which reports the income premium over a BA will likely hold true, or mostly true, in a future in which the economic value of a BA will also be devalued. The two sides simply disagree as to whether, as compared to previous legal recessions, this legal recession is (a) worse relative to the overall recession; and (b) different in nature.

I’m going to discuss each of these issues, and also say a few words about the policy implications of (my take on) this study. But before I do, please allow me a lengthy aside to raise a logically prior question: To what extent are Alice and Bob — tomorrow’s potential future 1Ls — likely to be competent to make good forecasts about where they likely will fall on the income distribution?

A. Law School Applications and the Dunning-Kruger Effect

I think that an intuition that Alice and Bob have a tendency to be over-optimistic drives a good chunk of the rage at ‘The Economic Value of a Law Degree’. There may be something to this fear, especially in a steep recession like this one. Right now the number of law school applicants has dropped some 13% over last year. I don’t have the latest data after what appears to have been a late surge in applicants, but back when the decline looked to be more than 20% what we were hearing was that the decline was especially steep at the high end of the credential distribution, that is among students with good grades and high LSATs.

If that’s true, it does raise the issue whether many Alices and Bobs might have a systematic vulnerability to the Dunning-Kruger effect.

The Dunning-Kruger effect is a cognitive bias in which “people reach erroneous conclusions and make unfortunate choices but their incompetence robs them of the metacognitive ability to realize it”. The unskilled therefore suffer from illusory superiority, rating their own ability as above average, much higher than in actuality; by contrast the highly skilled underrate their abilities, suffering from illusory inferiority. This leads to a perverse result where less competent people will rate their own ability higher than more competent people. It also explains why actual competence may weaken self-confidence because competent individuals falsely assume that others have an equivalent understanding. “Thus, the miscalibration of the incompetent stems from an error about the self, whereas the miscalibration of the highly competent stems from an error about others.”

If this is a real phenomenon, then in light of the current legal recession and the declining over-all prospects for law graduates we should (at least before the lessons of ‘The Economic Value of a Law Degree’ get internalized) expect more of the best students to stop going to law school, out of an unjustified fear that they will not be the best, and won’t get the best-paid jobs. Meanwhile, given that there are still some good jobs at the top, even if fewer than before, less competent students will — incorrectly — believe they are the ones who will get the best grades and get the best jobs, and will continue to apply in large numbers. We might thus see a decline in the credentials of the national applicant pool as a whole, and especially in the applicant pool of non-elite law schools.

But wait, it gets worse: if the Dunning-Kruger effect is large in absolute terms, then the absence of the best students will open up some space at the top, and students who formerly were in the middle of the pack will now find themselves at the top of the curve, validating their erroneous beliefs as to their abilities. In short, Dunning-Kruger effect + grading on a curve = wrong signals.

In this view of the world, one of the most important real-world effects of ‘The Economic Value of a Law Degree’ may be to draw back the more able students — the ones who take the trouble to collect data on their prospects — into the applicant pool. Perversely, however, this will make the next cohort’s Alices and Bobs worse off, as now they will be competing on graduation with a stronger talent pool, thus perhaps diminishing their place in the outcome distribution.

I mention this first, before getting to any advice to Alice and Bob about how to use the data in ‘The Economic Value of a Law Degree,’ because the more I think about it, the more I suspect that the effect on the composition of the applicant pool, even more than its size, may be the most lasting effect of this paper: if this data holds up, smart students who do their homework carefully and would be strong applicants to law school may be less likely to be frightened off without reason. Conversely, poor students, who don’t read carefully, may misunderstand the results of this study and the attendant publicity and suffer from unreasonably optimistic estimates of their prospects. Both critics and supporters of the paper might be right since they are looking at different pieces of the elephant.

I don’t know for sure if any of this is even vaguely correct, and there’s certainly nothing in the data I’ve seen that would tell me. But it does illustrate some ways in which it’s hard for even a careful, risk-averse person to guess their future income stream.

B. ‘This Time Is Different’ — Or Not

Recall that one of the big issues that divides Simkovic & McIntyre from their critics is to what extent the current legal recession is due to something special about a change in the demand for legal services — the claim that this time really is different. Simkovic & McIntyre say that the past ought to predict the future, that they don’t think anything terribly special is going on. I can’t claim special expertise on this debate. For what little it’s worth, I tend to think we’re all in the soup together, and that globalization, outsourcing, not to mention robotics and applications of big data will reduce the premiums for many professions, including both law and medicine, but I can’t point to anything that much stronger than a hunch based on current data and what I read about developments in technology. You might reasonably reply that it’s all very well for me to say that, but isn’t it darned convenient to my own economic interests, and a neat way to avoid the discomfort I’d have to take on if I thought law school really was a scam?

To those who say that the last couple of years have been particularly ugly for law graduates compared to past years, I can only say that my understanding is they were also pretty ugly for college graduates. According to 2012 BLS data,

About 1.5 million, or 53.6 percent, of bachelor’s degree-holders under the age of 25 last year were jobless or underemployed, the highest share in at least 11 years. In 2000, the share was at a low of 41 percent, before the dot-com bust erased job gains for college graduates in the telecommunications and IT fields.

Out of the 1.5 million who languished in the job market, about half were underemployed, an increase from the previous year.

The lawyer numbers are bad, but these are, if anything, worse. Things are really bad out there for a lot of people. (This is why macroeconomic policy is so important.) It may be that those of us involved in the production of new lawyers and the reproduction of hierarchy are, or tend to know more of, the ones with a new JD. As we will see, that doesn’t by any stretch mean that every law degree is a good investment. But it does mean that the problems of buyer’s justified remorse may not be due in the main to something about the changes in the nature of the demand for legal services.

Coming In Part 2: How Much Should a Prudent Student Be Willing to Pay for a JD?

1. In his latest comment, Brian Tamanaha suggests a different divide: whether one focuses on the long term, in which maybe a JD isn’t a bad investment, or the short term, in which recent graduates facing huge debt repayment and lower salaries than they hoped for face a serious cash crunch. [↩]

While I’ve been finishing some work and riding on airplanes the debate over Michael Simkovic and Frank McIntyre’s paper The Economic Value of a Law Degree has been raging in the blogosphere. There is an element by which the parties are talking past each other but I think there is light as well as heat.

Here are links to the counterpunching highlights, more or less in order:

• Michael Simkovic, Brian Tamanaha Says We Should Look at the Below Average Outcomes (And We Did)
• Brian Tamanaha, , Brian Tamanaha, How “The Million Dollar Law Degree” Study Systematically Overstates Value: Three Choices that Skewed the Results
• Michael Simkovic, Brian Tamanaha’s Straw Men (Overview)
• Brian Tamanaha, , How the “Million Dollar Law Degree” Study Understates Risk (Part I)
• Michael Simkovic, Brian Tamanaha’s Straw Men (Part 1): Why we used SIPP data from 1996 to 2011
• Brian Tamanaha, Why the “Million Dollar Law Degree” Study Fails (Final Post)
• Frank Pasquale, Why Tamanaha’s “This Time is Different” Critique Fails

[Update (7/25/13): Add Brian Tamanaha’s Straw Men (Part 2): Who’s Cherry Picking? to the list.]

Below I want to correct an error in one of my earlier posts. In my next posting on this topic I plan to offer a few relatively timid thoughts describing what I’ve taken away from the debate so far.

Here’s the correction: In Simkovic and McIntyre Respond to Critics (Updated), I wrote, regarding people who don’t enjoy law school, that,

In some cases the cost of three years of hell may not justify a lifetime increase in net income of, say, $350K. After all, over a 40-year career, that’s an average of just $8,740/year…and indeed most of it comes in the out years, not early when you are struggling to repay debt.

This was a mistake because it confused present discounted dollars (the currency of the Simkovic & McIntyre paper) with nominal dollars. In nominal dollars the actual premiums predicted by their paper would be much higher, particularly in the later years since lawyers typically make more when they are senior and the out years are more heavily deflated in a NPV calculation.

On the other hand, I should also have emphasized that the study compares a JD to a no-JD, ie BA-only, world. It doesn’t compare a JD to an MBA or other advanced degree. Even taking all the study’s assumptions on board, the premium for alternate degrees might be higher than a JD, or in any case not as big as the JD/BA.